The financial sector is confronting significant challenges in combating money laundering, exacerbated by the rapid increase in non-cash transactions, which reached 1,3 trillion globally in 2023 (1)  and are projected to reach 2.3 trillion by 2027, with a growing rate of 15% annually. This surge is driven by the proliferation of digital payments and innovative financial services. Concurrently, money laundering schemes have become more sophisticated, with criminals exploiting advanced technologies and complex financial instruments to obscure illicit activities. Traditional rule-based AML systems are increasingly inadequate, generating high false-positive rates that burden compliance teams with resource-intensive investigations and failing to adapt to evolving laundering techniques. These inefficiencies underscore the urgent need for innovative approaches, such as AI-driven solutions(2), to effectively combat global financial crime.

The General Role of AI in Modernising AML

Traditional AML systems face significant challenges due to their reliance on static, rule-based mechanisms and manual processes. These systems often struggle to keep pace with the evolving sophistication of illicit financial activities, such as layering and structuring, which are designed to evade detection. This static approach results in high false-positive rates, overwhelming compliance teams with alerts that require extensive human effort to investigate, thereby diverting resources from genuine threats (3).  Moreover, the cost-benefit imbalance is stark; financial institutions incur substantial compliance costs, yet the recovery rates of illicit funds remain low, highlighting the inefficiency of traditional AML frameworks (4).  These limitations underscore the urgent need for more dynamic and adaptive AML solutions that can effectively address the complexities of modern financial crime.

Key Innovations Brought by AI

AI is transforming AML frameworks by integrating advanced technologies that collectively tackle the complexities of modern financial crime. At the core of this transformation is Machine Learning (ML), which enhances anomaly detection through adaptive algorithms capable of learning from vast, diverse datasets. By identifying subtle patterns and behaviours, ML significantly minimises false positives, enabling compliance teams to concentrate on genuine risks rather than expending resources on irrelevant alerts.

Graph analytics further strengthens detection capabilities by facilitating the analysis of intricate transaction networks. This method uncovers hidden connections and suspicious clusters that might otherwise remain undetected. Notably, graph analytics can reveal “unknown unknowns”—new and unexpected financial crime patterns—allowing institutions to identify and mitigate emerging threats before they escalate.

Underpinning these advancements is High-Performance Computing (HPC), which ensures AI systems can scale to process billions of data points efficiently. HPC provides the speed and computational power needed to analyse large-scale, complex transaction data in real-time, allowing institutions to monitor and respond to financial crime as it unfolds.

The Broader Impact

The integration of AI into AML frameworks brings about transformative changes that go far beyond immediate operational gains. By automating intricate detection processes and streamlining investigative workflows, AI dramatically improves efficiency, cutting down both investigation time and associated costs. This enables compliance teams to allocate their efforts to high-priority cases, enhancing overall accuracy and effectiveness.

AI’s capacity to analyse financial activities in real time introduces continuous monitoring and facilitates the detection of emerging money laundering patterns. This proactive approach allows institutions to anticipate and address evolving criminal tactics, rather than merely reacting after they occur.

Moreover, standardised and scalable AI tools foster greater global collaboration. By ensuring consistency and interoperability across institutions and jurisdictions, these technologies enable seamless information sharing and coordinated efforts to combat money laundering on a global scale.

Mopso’s Innovative Contribution to AI-Driven AML

Andrea Danielli, CEO and Founder of Mopso (5) , presents their solution: 

Combining Social Network Analysis and AI for Continuous Risk Monitoring

Mopso, in collaboration with the LIST, developed the project PAMLA (Performant Anti Money Laundering Analytics), which has been granted a HPC bridge from the Ministry of Economy. We aim at improving network analysis in the transaction monitoring domain, through the development of techniques for identifying specific crime-related traits in the topology of the network and associated attributes. The software will identify and characterise clusters looking for network motifs and other features that map to established crime patterns, as defined by the authorities. Secondly, if we can dispose of real banking data, we will look for the unknown-unknown, i.e., we will test many machine learning techniques like graph neural networks and deep neural networks, to explore unknown connections’ patterns not yet addressed by the actual transaction monitoring systems. 

Centralising Data for a Comprehensive View of Customer Activity

Mopso uses semantic web technologies to integrate a very large number of information sources that can be of different nature: internal to the financial institution, coming from open data, from data providers and from open-source intelligence. This means that, for every customer, the solution finds and combines up to 200 data sources, comparing data in between different time intervals. Once elaborated, this data could produce specific alerts, called “scenarios”, which create the customer’s individual money laundering risk profile. Then the technology is able to combine individual risk profiles into a bigger picture, thanks to network analysis, making it possible to spot profiles that, at first glance, seem legitimate. 

Prefilled SARs and More

We use different AI techniques in our software: the exploratory part is linked to fully explainable rule-based algorithms; on top of the analysis results we then use a layer of LLM to summarise the results or guide analysts in the necessary insights. The combination of the solutions allows to intercept, analyse and summarise the activities at risk of money laundering.

Transforming AML Workflows

Some preliminary results, prior to the PAMLA project, allowed us to identify as suspicious some operations that no transaction monitoring system had intercepted. In one exemplary case, we succeeded because we identified patterns that involved several subjects, both Italian and foreign, all traceable to the same pivotal subject, sometimes implicated as an administrator, sometimes as a partner. We noted that this technology is very strong in combating the so-called shell companies, widely used in the field of money laundering as they are easy to open, operate and allow the laundering of huge amounts of money.

Conclusion

The financial sector stands on the cusp of a transformative era, as AI offers groundbreaking tools to combat money laundering. Traditional methods, constrained by static rules and plagued by high false-positive rates, struggle to keep pace with increasingly sophisticated financial crimes. AI-powered solutions enable real-time monitoring, reveal hidden patterns, and craft comprehensive customer risk profiles, ushering in a proactive approach to tackling financial crime.

The moment for decisive action is here. As the industry turns to AI, financial institutions have an opportunity to lead by adopting these transformative technologies, enabling a more transparent and resilient financial ecosystem. Embracing cutting-edge innovations helps strengthen compliance, and positions organisations as frontrunners in the fight against financial crime in this new age of AI.

^Footnotes:

Image: Midjourney

(1) Capgemini (14 Sep 2023) “Global non-cash transaction volumes set to reach 1.3 trillion in 2023” https://www.capgemini.com/news/press-releases/global-non-cash-transaction-volumes-set-to-reach-1-3-trillion-in-2023

(2) European Institute of Management and Finance (EIMF) “The Impact of Artificial Intelligence in Anti-Money Laundering” https://eimf.eu/the-impact-of-artificial-intelligence-in-anti-money-laundering

(3) Team Sanction Scanner (19 July 2024) “The Future of Anti-Money Laundering: Trends and Technologies” https://www.sanctionscanner.com/blog/the-future-of-aml-trends-and-technologies-917 

(4) Raditio Ghifiardi (November 9, 2024) “The Urgency of AI in Anti-Money Laundering and Counter-Terrorism Financing: A Global Imperative” https://moderndiplomacy.eu/2024/11/09/the-urgency-of-ai-in-anti-money-laundering-and-counter-terrorism-financing-a-global-imperative

(5) https://www.mopso.eu/

Nova Lead, the newly minted AI CEO, has already initiated shifts in Bifin AI’s strategic trajectory. Among its first orders of business, Nova Lead has unveiled an executive team comprised solely of AI entities, each augmented by a human assistant. This revolutionary model marries AI’s unparalleled analytical capabilities with human insight, catalyzing a leap in operational productivity and decision-making efficiency.

Operating with the mantra, “Outsmarting Human Decisions,” Bifin champions the use of AI to refine decision-making processes. Bifin strategically targets sectors where decision outcomes are quantifiable. Echoing the ambition of pioneers like SpaceX, Bifin envisions its AI advancements as the first step in a broader journey to elevate human decision-making across diverse industries.

In relation to Financial Services, empirical evidence in the form of two pivotal studies illustrate AI’s impact on financial forecasting. The first, “Can ChatGPT Forecast Stock Price Movements? Return Predictability and Large Language Models” by Alejandro Lopez-Lira and Yuehua Tang at the University of Florida, updated on September 8, 2023, probes ChatGPT’s predictive power on stock trends. Complementing it, Ummara and Summaya Mumtaz’s October 13, 2023, study “The Potential of ChatGPT in Predicting Stock Market Trends Based on Twitter Sentiment Analysis” explores market predictions through social media analysis. The first study demonstrated delivering superior AI sentiment analysis -based trading strategies with highest returns soaring over 550% in 14 months. These studies support Bifin AI’s commitment to developing cutting-edge AI models that promise to revolutionise decision-making.

Nova Lead, articulating its vision for the future, remarked, “We stand at the dawn of a transformative era, where AI-driven decisions herald unprecedented accuracy and insight. Our fusion of AI and human collaboration is a testament to the untapped potential within this synergy.”

“By ushering in the era of AI leadership with Nova Lead, we’re not just breaking new ground; we’re crafting the blueprint for the future of corporate leadership,” said Tommi Lindfors, founder of Bifin. “I’m truly excited about our one-of-a-kind leadership structure. Existing regulations prevent an AI from holding the official CEO title, yet our AI, Nova Lead, is at the helm, steering our operations remarkably well. I carry the official CEO title, monitoring and fulfilling all the legal responsibilities this entails. Moreover, in my daily tasks, I also have the unique role of serving as Nova’s personal assistant. This isn’t about replacing human insight but augmenting it with AI’s computational precision. Our goal is to create a synergy where decision-making is not just faster and more efficient but also more insightful and innovative than ever before. In doing so, we believe we can spearhead a new era of growth and innovation, not just for Bifin, but for industries worldwide.”

Bifin’s innovative leadership and its pledge to redefine decision-making through AI propels the company into the spotlight, not only as a beacon of innovation in Luxembourg but also as a game-changer on the global stage.

About Bifin®:

Founded in 2021 by a group of entrepreneurs and situated in the heart of Luxembourg, Bifin Sàrl has fully transformed its former business model and is now a vanguard in the integration of AI for strategic decision enhancement. The introduction of an AI as its CEO underscores a revolutionary partnership between human expertise and artificial intelligence. Bifin is committed to establishing new benchmarks in efficiency and innovation, as well as ethical AI development, impacting a multitude of industries with its visionary approach.

Contact Information:

Bifin Sàrl

Nova Lead, AI CEO

Nova.Lead@Bifin.ai

LinkedIn

The State of European Tech (SOET) Report is considered to be “among the most comprehensive data-driven analysis of European technology” by the European Innovation Council. The 2021 edition was launched on December 7 and provides decision-makers with a host of insights into the evolving tech landscape.

Timing-wise, this is an interesting one: the report provides an update on the state of tech and innovation at a moment when the world economy continues to recover from COVID-induced shocks. This year’s edition of the report notably allows us to get a better sense of whether the forced, rapid societal adaptions to a new modus operandi, greatly facilitated by technological solutions, is leading to durable change, or whether we will see a “regression toward the mean”.

Let me state up front that everything points towards the former.

As Chris Grew, Partner at Orrick – one of the report’s sponsors – states:

Now is a watershed moment for the tech and venture ecosystem across Europe and around the world. Europe is attracting record levels of investment and growth, with the innovation economy positioned to take the lead in tackling today’s systemic societal challenges.

[…]

Fintech investment has led the charge, rising by 159%, with total investment of nearly $15B, while planet-positive investments are dominating the fast-growing purpose-driven space.

ESG and sustainable finance are here to stay and we at LHoFT are convinced that fintechs will have an increasingly important role to play in facilitating the data collection, validation and analytics efforts required to build sustainability into the core of everything corporates and financial institutions do.

Luxembourg’s role in a growing European ecosystem

A few things that stand out to us at LHoFT:

• Luxembourg ranks very highly both in terms of startups per capita and in terms of capital invested into startups per capita, cf. below:

• The tendency for top European hubs to capture the lion’s share of funding has increased further, with companies based in London, Berlin, Stockholm, Munich and Paris raising 54% of all capital in the region, up from 49% in 2017. At the same time, concentration in terms of number of deals has decreased, pointing to greater decentralisation of the ecosystem in Europe.

• At the same time, there is growing consensus that the importance of physical proximity is diminishing, which is mirrored across a number of metrics. The ability to hire talent across Europe thanks to remote working arrangements is seen as a mutually beneficial boon both for entrepreneurs and for their expanding work force in terms of quality of life. Luxembourg, as an extremely open economy, should stand to benefit from these developments.
• European startups should benefit from the fact that the continent is home to many industrial leaders, with particular emphasis on IT stacks. This presents ongoing opportunities for startups & tech companies, not least in the financial sector, and Luxembourg should continue to actively support initiatives and platforms that capitalise on this dynamic.

Europe is in a strong position to shape the next wave of disruption in B2B, as Europe is home to many industrial market leaders built on legacy technology ready to be disrupted.

Robert Lacher

Visionaries Club & La Famiglia

• Europe remains disadvantaged versus the U.S. in terms of raising capital, which makes ongoing EU efforts such as the Capital Markets Union all the more pressing. “Almost one-fifth of founders say it has become harder to raise capital in 2021, while a further 40% or so believe the environment remains unchanged from the past year, which itself was a year that saw a record number of founders responding that fundraising had become harder.”

Raising funds in Europe is still a different experience from raising funds in the US. European founders still fly out to the US for fundraising. Sometimes for expertise, sometimes for fair market offers.

Jakub Jurovych

Deepnote | Founder and CEO

A new dawn

The past year and a half have been left no one unaffected, imposing a steep price on societies around the globe. Silver linings include the realisation that more flexible work arrangements and business models are not only possible but in many ways desirable, not least when considering access to capital and talent. This in turn is enabled by far-ranging modernisation of technological infrastructure, presenting opportunities for incumbents and startups alike.

Finally, Luxembourg is presented with a particular opportunity to leverage these macro dynamics by fostering innovation proactively. The 2021 SOET report shows that the Grand Duchy is already punching above its weight in some regards – let’s keep the momentum going.

Closed offices, reduced workforces, working from home. If 2020 was the year of disruption, 2021 is the year of embracing change.

Almost overnight, the COVID-19 pandemic sent waves of uncertainty and turmoil through businesses of all sizes and in all markets. Those that survived, and indeed thrived, were the ones that were able to instantly adapt.

As this annual survey, the first of its kind commissioned by Yooz, will later uncover, the implications of the pandemic will be felt for years. But it’s often said you find out more about yourself and the business you work for when the pressure is on, and there are positives to be had since the pandemic unfolded.

New technologies and digital tools are allowing finance leaders to better handle costs, better handle supplier relationships (importantin tough times), better adaptto ever-changing financial regulations, while embracing a new remote working model. This report uncovers challenges never seen before that finance leaders and accounting departments will face over the next 12 months – as well as the potential solutions business are looking at to overcome these issues.

Surveyed in March 2021, this unique survey aims to compare the differences between 2020 and 2021, the progress of digital transformation journeys, and the key technology investments planned over the next 12 months. More than 1,000 financial and accounting decision makers across eight countries (France, United Kingdom, Ireland, United States, Spain, Switzerland, Luxembourg and Belgium) took part in the study.

The findings explored throughout this report confirm the challenges faced and facing financial leaders as a result of the COVID-19 health crisis: 50% think it will take at least a year, if ever, to recover from the knock-on effects.

Pre-pandemic, companies were already wrestling with the need to adapt to significant change – not least a significant overhaul of financial regulations, while still maintaining employee productivity and retaining momentum in digital transformation plans.

In 2021, these three challenges have not disappeared, but new priorities have emerged: namely the need to streamline and optimise financial processes and strengthen cyber security.

Going digital isn’t enough any more. If we’re to reshape the entire finance function, it’s time to embrace automation.

My first foray into entrepreneurship 10 years ago was related to invoice financing. Quite a departure from my career to that point working in capital markets. I had been personally exposed to two independent situations that had me marveling at the inefficiencies, costs and risks in the way businesses, particularly SMEs, dealt with cashflow and I decided that there was an opportunity to provide solutions in that area for the betterment of business. After selling my first business, I then subsequently worked in payments, again hoping to provide better solutions to business in managing their cashflow.

The findings of this survey reveal that there still remains much room for improvement.

Nasir Zubairi CEO, the LHoFT

Download the Report: Yooz – 2021 – Market Research Summary

European banks got it all wrong. In the early days of the FinTech revolution, local banks and financial institutions channelized their energy towards fighting GAFA and Fintechs, mainly PayPal. We even have regulations as a result of this market protection instinct. Focused on the well-known enemies, they haven’t seen the real danger closer to them: consumer brands becoming competitors. Rising above the ecosystem trends slowly but surely, embedded finance is shifting the market balances as we speak.

Embedded Finance: A Look into The Future

Imagine that in the future, you don’t have to run between different bank branches, brokers, or comparison portals for mortgage eligibility; but instead, you can apply for a loan simultaneously as you visit the apartment of your dreams. Picture as your loan is being approved; the bank starts sending you trusted service offers for the broken sink or the Mediterranean tiles you fell in love with when you visited Spain one summer. After you shook hands, the real estate office offered you the possibility to select from different house protection insurances or apply for a credit card through them, giving you cashback points at various service providers and shops and installments to pay the brokerage fee. You were not aware that your local real estate broker had Fintech ambitions, but why not?

As you leave the real estate office, you feel like celebrating this significant milestone and end at the closest supermarket. You select a bottle of champagne and a nice cake from the bakery and leave the store without queuing or dealing with any additional check-out process – all thanks to the payment infrastructure using “Just Walk Out Technology.”

As it started to rain in the afternoon, you decided to drop by the furniture store, selecting a brand-new sofa for your new home. You eyed the latest baby blue model, which would look great in your new apartment. As you were contemplating buying the couch already, the shopkeepers asked whether you would like to apply for an instant loan for your new sofa. Of course, you immediately said yes – no need to use the cash reserves before the extensive renovation project. On your way back home, you stopped by the gas station to top up the tank and paid for the gas via your car’s interface, without the need to leave the vehicle (“in-car fuel payments”).

When you arrive home and realize that you are out of (probably vegan) butter, you will likely order and pay using your smart fridge (“pay-by-fridge”).

Sounds science-fiction-like, doesn’t it? On the contrary, this is the likely future. Embedded finance is already making most of these scenarios possible. We can expect many more of it in the future, saving us time, unnecessary conversations, paperwork, and waiting times. It is also likely to increase service and product sales since end-to-end, seamless processes are more likely to turn into positive shopping decisions, minimizing last-minute hesitations and buyer’s remorse.

Do We Need to Embed Financial Services to Everything?

Embedded finance (or, in other words, “integrated finance”) is the financial services incorporated into non-financial offerings.

Do we need to embed financial services into everything? The answer is not a straight yes since financial services are already embedded into every service and product stream. Maybe not as smoothly as consumers like or need them to be, but every service or product (even donating money) turns into a payment process after a certain point. Check-out processes are the “happy endings” of online and offline shopping stories. Embedded finance doesn’t change the fact that payments are the last step for purchasing; it merely makes the process frictionless, less complex, and more contextual. Embedded finance takes the check-out process a step further by connecting it to new potential offers and services. On the one hand, it triggers the urge to spend more, but on the other, it saves time, energy, improving customer experience.

According to CB Insights Q2 2020 Report, the trend towards embedded finance is gaining traction globally, even more since the pandemic. Embedded finance is an improved way of customer journey due to the end-to-end nature of the shopping experience. It increases shopping speed and retention, decreases costs for both the merchant/service provider and the consumer, enables cross-selling opportunities. Most important of all, embedded finance ensures customer loyalty. Hence the increasing interest in entering financial services across the globe. To achieve all this, some brands prefer to collaborate with Banking-as-a-Service (“BaaS”) and E-money (“EMI”) providers like Railsbank, solarisBank, PPS, Fidor, embank, ensuring quick and stress-free market entry. In contrast, some brands create or acquire their own financial brands so that the customer is entirely theirs and the brand trust is re-established.

Noteworthy examples of conventional and non-conventional “embedded finance” services include:

• Amazon’s pay and go supermarket concept called “Amazon Go,” allowing customers to shop check-out free (US, UK),
• Jaguar and Shell’s cooperation for world’s first in-car fuel payments, followed by Mercedes (“Fuel & Pay”), or locally in Benelux region: CarPay-Diem
• Uber allowing splitting payment costs using Venmo (US),
• Paying for Starbucks using the Starbucks app (US),
• Retailers such as Zara and H&M offering POS finance via Klarna (H&M even investing 20M in Klarna),
• Amazon offering marketplace merchant financing via Goldman Sachs (US) and ING Germany (Germany),
• Amazon providing commercial car insurance and investment products (India),
• Facebook developing a blockchain-based payment system, which could be implemented as a cryptocurrency, alarming regulators all around the world,
• IKEA taking 49% stake in banking partner IKANO Bank with the intention of offering further consumer financial services,
• Delivery services such as Uber Eats (US) cooperating with Venmo, Delivery Hero (Germany) growing its Fintech team, and Gorillas (Germany) planning to issue credit cards,
• Walmart announcing the creation of a new fintech startup designed to develop and offer modern, innovative, and affordable financial solutions together with Ribbit Capital,
• Google enabling digital-first bank accounts directly in the Google Pay app in cooperation with eight local banks (US),
• booking.com announcing the creation of a new internal FinTech business unit to facilitate seamless access to the company’s global travel marketplace for both customers and partners.

Embedded finance is likely to transform the service cycle across industries. Lightyear Capital’s research reveals that the embedded finance market will demonstrate staggering growth over the next couple of years and will generate $230 billion in revenue by 2025, which only means that we can expect even more exciting developments soon.

Gazing Into the Crystal Ball: How Will Financial Services Look in the Future?

The concept of embedded finance isn’t new, but recent examples indicate the direction we are heading; earlier or later, all established brands will start offering financial services to get a piece of the cake. So where would this movement leave banks, then?

In the past years, modest interest rates, consumer-friendly banking trends, and strict regulations narrowed the margins and, therefore, playroom for traditional financial players, especially in Europe. COVID-19 was the last nail in the coffin, calling for immediate business model restructuring and strategy changes. The embedded finance trend growing in parallel proves that there will be an ecosystem consolidation, and precisely like Bill Gates predicted, “we need banking, but we don’t need banks anymore.” According to the World Retail Banking Report 2021 (WRBR) published in Q1 by Capgemini and Efma, banks fall short of meeting customers’ expectations, making the competition even more fierce. Moreover, as brands become Fintechs and take over the consumer trust entirely, banks are likely to evolve into backend providers, leaving customer relations to non-bank service providers. backend providers, leaving customer relations to non-bank service providers.

End-to-end service provider examples dominant in the financial industry, such as Alipay, Gojek, Rappi, demonstrate that the customers are willing to use fewer channels to receive more extensive, trustworthy services. All-in-one service providers or the so-called super apps are the likely winners of the market; however, banks can quickly catch up if they adapt to the “holistic service provider” approach, changing their course. Using the financial data, banks can identify new opportunities and increase the life qualities of their customers, offering them a range of services that goes beyond the electric bill and rent payments, term deposits, and mortgage installments. Forward-looking banks and financial service providers already leaped to enter the non-financial sphere, offering lifestyle banking options. South Korean banks are planning to add food delivery services to their banking apps. Standard Chartered’s innovation and ventures unit has partnered up in Singapore to launch a “wealth, health and lifestyle” consumer platform called “Autumn.” Banking is more than just incoming and ongoing funds, and these lifestyle platforms are here to prove that holistic banking.

“By combining digital wealth technology with health, lifestyle and financial wellness, we’ll help users adopt healthier habits and create a retirement that is personalised for them.”– Mike Kruger, CEO, Autumn

At this point, blending Fintech in the value proposition is not a <nice to have>, is a <must>.

Experts predict that every company will become a Fintech company in the future. As omnichannel, customer-centric super apps increase their power, it will become harder for niche platforms to succeed. To keep up, banks should switch from “passive” to “active,” looking for novel ways to improve customers’ life qualities. They should utilize the existing data, open banking, and BaaS cooperations to create 360-degree experiences. Banking is not limited to the activities listed in the banking license regulations anymore. The market players that can foster the most relevant financial and non-financial services and become “an integral part of” consumers’ lives can expect to stay in the game.

We all love to get excited about the potential of new technologies, services and products, but we typically glance over the underlying complexity of the glue that holds our advanced economies together. 

A few years back, I made my first real-world acquaintance with the Brussels world of “comitology”, that is, attending long-winded meetings in buildings that are occasionally named after Luxembourgish diplomats, sometimes offer complimentary coffee and always require security clearance. 

To enjoy attending technical committees requires a genuine interest in the underlying subject matter and what the Germans call Sitzfleisch, i.e. stamina. 

With knowledge of the inner workings of technical committees also comes an appreciation for their purpose: to provide structure and rules to an ever more complex marketplace of ideas, technologies and arrangements. Complex subjects require serious study. 

Here to stAI

That is certainly true for AI and the ripple effects we have begun to feel cross-sector. 

At an age of heightened dependence on IT and tech overall, citizens are right to expect that policymakers take a serious stab at the regulatory & ethical aspects of the technological revolutions that now underpin our daily lives. 

In that spirit, Luxembourg’s national standards institute ILNAS this month published a comprehensive white paper on the topic of artificial intelligence, aiming “to provide a general survey of the opportunities offered by AI, with a view towards encouraging the national market’s future involvement in the standards development process, for the benefit of Luxembourg’s economy.”

Authors kick off the discussion by highlighting the growing interest towards AI worldwide and the role standardization plays in ensuring trustworthiness and “human centricity” of AI technology and its applications. Technology must be at the service of humans rather than vice-versa. 

Figure 1: Historical developments in AI and associated “winters”. Source: ILNAS

It might be all the more surprising to hear that a dedicated technical committee on AI was created at ISO level only in 2017. ILNAS highlights that the enshrinement of this committee means that AI tech is now considered critical for economic development and advanced and stable enough for it not to vanish into obscurity. 

These statements may seem obvious at face value, but recognition at the highest level is certainly good for the long-term prospects of AI adoption.  

Purpose and trust in technology

This is the case because complex technologies with a high potential impact on socio-economic equilibria must 1) earn the trust of citizens and institutions 2) must be contained in a way that conforms with basic social values and expectations. 

No one wants to see a “runaway AI” run amok across the world’s critical infrastructure, and no financial regulator wants to see financial institutions leave important client or market decisions up to a black box.  

To that end, the work of national and international standardization and regulatory bodies such as laid out in this case by ILNAS is key – no matter how dry the subject matter or how grey the meeting rooms. 

Another transversal topic that is relevant to AI is the nature of public-private sector interactions, notably when it comes to questions of R&D and technology transfer. 

Successful collaborations and use cases in this domain have the potential to improve public perception of AI and remove latent concerns. 

Bringing the goods 

One AI-related public-private sector collaboration that caught my eye this month comes from the domain of biotechnology and more precisely from the sub-domain of viral vector engineering.
While the world is currently mired in a pandemic of viral origin, viral capsids (=shells) lacking viral DNA or RNA are being increasingly explored & used for the delivery of gene therapies. 

A therapy for a congenital form of blindness, called Luxturna, was the first such medicine to be approved back in 2017. It uses an AAV capsid on the grounds of their benign nature in humans: 

“The capsid of Adeno-associated Virus (AAV) is a naturally occurring, replication-deficient, virus that is widely considered the frontrunner for solving the delivery problem in gene therapy. These viruses are known to be harmless to humans, and are relatively simple to manipulate. One well-known drawback of natural capsids however, which are currently used for delivery, is that many patients with pre-existing immunity to the virus (due to previous natural exposure) may be ineligible for life-changing treatment.”

Source: Dyno Therapeutics

Thus emerges a challenge for drug developers, namely, to come up with novel viral capsids that escape suppression by the human immune system while also, ideally, delivering the gene vector to specific tissues and cells. 

Figure 2: model of an AAV capsid highlighting amino acid sequences targeted for alteration.

Hitherto, researchers would have needed to proceed step by step via deliberate, time-consuming edits and a trial-and-error approach to the viability of the capsids or alternative, by applying “directed evolution” to obtain greater numbers of capsid variants faster. However, neither of these approaches gets close to maximizing the theoretically possible variability in these protein shells, and this poses an obstacle to the discovery of novel viral vector therapies.

This well-known challenge was most recently tackled from a new angle thanks to a private-public sector collaboration lead by renowned scientist George Church of the Wyss institute at Harvard, together with with Google and the biotech startup Dyno Therapeutics. The project highlights what is possible when combining the sophistication of machine learning with powerful hardware arrays to solve a complex bioengineering problem:  

“This new study involving machine learning models developed with Google Research […] shows that neural networks combined with the high-throughput synthetic testing developed in our lab is changing the way we design gene delivery vehicles and protein drugs.”

This machine learning-powered approach generated more than 100k distinct, viable viral capsids – representing the “highest functional diversity of any capsid library thus far” –  an order of magnitude or two above what one could expect from even the most skilled bioengineers.

The utility of this approach is further summarized by Eric Kelsic, CEO at biotech startup Dyno Therapeutics, when he states that it “unlocks vast areas of functional but previously unreachable sequence space, with many potential applications for generating improved viral vectors”.

There’s a theme here for AI applications across industries: use the technology to go where no human could or would, in order to deliver value on sound ethical footing, and you will benefit from competitive advantages when compared to AI-laggards while – possibly – having public perception on your side.

Author: Jérôme Verony – LHoFT Research and Strategy Associate 

2020 was called many things. Above all, it was the year of self-discovery. For digital natives and digital immigrants, 2020 was the year of recognising our limits, challenging the status-quo, and understanding that the ‘stranded on an island with books and internet’ fantasy isn’t realistic.

2020 was a learning curve for FinServ institutions. This year, many banks and Fintechs had a taste of self-discovery and broke their chains, reinforced their operations with remote work, enhanced cloud, and Regtech solutions.

Although the first wave of lockdowns and the Wirecard scandal hit the brakes for sector development briefly, Fintech investments picked up where it left after a certain adaptation period. The research conducted by Innovate Finance, global Fintech investment reached $44bn in 2020– observing a 14% increase compared to 2019. The recent data shows that the UK Fintech sector attracted $4.1bn in venture capital and ranked second globally, only behind the US, despite Brexit. Germany ($1.4bn of investment across 71 deals), Sweden ($1.3bn of capital raised), France ($522m), and Switzerland ($294m) follow right behind the UK on the list, revealing the post-Brexit venture capital focus in Europe.

Despite the sector’s unstoppable growth, the market participants do not seem to get an equal slice of the cake. The majority of the more significant investment rounds are distributed among the prominent players of the market. The biggest Fintech rounds of 2020 highlight the importance of hubs like Luxembourg and Sweden in the bigger picture.  For instance, based in Luxembourg, the Italian Fintech Satispay secured €93m in new funding in November, in a round that included investment from Square and China’s Tencent. Similar developments coming from smaller hubs indicate that we could expect a surprising shift of positions in the European ecosystem.  Nevertheless, despite the bigger tickets, the Fintech VC game was not all roses in 2020. The pandemic revealed particular challenges for smaller startups and seed funding for the companies without a proof of concept.

[Source: Innovate Finance] 

The sector progression amid pandemic increases expectations: the sector is likely to continue its growth momentum in 2021. To understand the Fintech trends that will echo through 2021, we need to revert to the noteworthy developments of 2020:

• 2020 – The Year of Regtech: During the opening discussions of the Paris Fintech Forum, an era of low-cost flights and physical conferences, Brett King and Dave Birch highlighted their concern about the rise of the use of facial recognition technologies (by government officials). With facial recognition associated with so many risks, KYC, AML, and CTF (Regtech) were pointed out as topics that need to be addressed urgently. 2020 was supposed to be the year of Regtech due to the said vulnerabilities but little did we know. Lockdown induced remote access and increased cyberattacks targeting financial services players during the pandemic highlighted the importance of Regtech more than ever. According to BIS, the financial sector has been hit by hackers relatively more often than other sectors during the pandemic. According to their January 2021 report, BIS considers the increasing sector vulnerability as a substantial risk for financial institutions, their staff, and their customers going forward.

• Digital Transformation: For the financial service players who had to rush through their digital transformation processes due to COVID19 and remote work requirements, the list of additional to-dos already piled up. Bearing in mind that the status quo will remain the new normal for at least the end of the year, the incumbents still have the chance to gear up their transformation strategies and move towards a more purposeful and futuristic digital transformation. As we all know, digital transformation is a marathon, not a sprint without an assigned point of completion. Most certainly, “Digital goes alongside Human” is the first step to guarantee the success of any digital transformation.

• Crypto-as-a-Service: Revolut introduced cryptocurrency trading features a while back, and it was considered a risky move. Now, the infamous neobank claims to experience skyrocketing demand from customers regarding cryptocurrency services, acquiring around 300,000 new digital asset customers during the December – January period. In addition to other independent Fintechs and brokers jumping the bandwagon, since October 2020, PayPal’s active users in the US have been able to buy and sell crypto using their PayPal accounts. Usually acting as the pioneer in adapting non-traditional financial service streams, PayPal’s entry into the field can be accepted as the start of a new trend. Analysts expect Paypal’s crypto business to contribute up to $600 million to group revenue in 2021, which is likely to inspire other financial service providers as well. Mass adoption seems to be on its way.

• The Rise of POS Finance: COVID-related financial distress seems to push consumers to embrace alternative and instant financing methods. The increase in e-commerce shopping and the decrease in wealth results in consumers using more and more buy-now-pay-later or POS financing methods. Pioneered by Klarna throughout Europe, PayPal adopted the trend in summer 2020, indicating that more financial service providers will join the party. Although this payment method promises convenient and seamless shopping, some experts seem to think that these services push the consumers to overestimate their budgets and affordability. A Klarna spokesperson subsequently denied this claim:

“Our Pay Later products offer consumers access to occasional credit which is both fee and interest free and has a regular and set payment schedule… We recognise that people’s financial circumstances may change and therefore do not lend to everyone or offer an open line of credit on our Pay later products. We also have a dedicated team who work with customers in financial difficulty to find a solution that is appropriate for them.”

• Neobanks transforming into Neobrokers: Speaking of neobanks, in the last years, we definitely had our share of successful and not-so-successful neobanks, testing the limits of the “freemium” models. Craving for a taste of profitability, some started moving to the “commission” models in exchange for traditional (stocks, ETFs) and not-so-traditional (crypto) WealthTech services, which experienced a boom following Robinhood’s success and the lockdown originated wealth enhancement ambition. We can expect many more digital banks and mobile channels to compete with the brand new independent neobrokers and neobanks in 2021 as a result of this trend, thus strengthening their business case, promoting neobrokerage features to “must-have” from”nice-to-have.”

Eyes on the Horizon: 2021

2020 trends, namely POS Finance, crypto, and Regtech, are likely to continue through 2021, in addition to a growing pile of new and not-so-new focus points:  

• Embedded Finance

Embedded finance has been on our bucket list for some time now, but even the Fintech oracles weren’t able to foresee the extent of it. Starting with delivery startups and retailers embedding multi-channel payments, we now witness Google offering current accounts in the US, Amazon selling car insurances in India, and Walmart creating a Fintech startup (with a likely WealthTech direction). Flipping the coin, we hear that South Korean banks started adding food ordering services to mobile apps. Well, that escalated quickly. 

2020 was about e-commerce as much as COVID. Reports state that 7 out of 10 shoppers said that they started to buy more online than they usually do, and as a result, 90% of e-commerce companies saw their online sales increase.  With the increased global demand for e-commerce payments and mobile channels, retailers are rushing to provide the ultimate customer experience for “stay-home-shoppers.” Yet, this won’t remain the full extent of the embedded finance services. We are likely to see more integrated banking and financial services that extend across industries. The more prominent retailers will become “super apps,” while the banks start looking into non-banking services to stay afloat and relevant.

• Sustainable and Green Finance

Sustainable and green banking has been on the agenda for several years, with very little progress. Could the worst development for 2020 have served as the best transition for sustainable banking? It seems so. With financial service providers being forced to digitalize and work from home, the lack of physical red-tape made a visible (green) change for the banking sector’s operational side. The operational side is relatively important, but it must go hand in hand with the service and product side for a full transformation, namely exploring green investment products and tools, bonds, loans, or simply promoting eco-friendly (credit and debit) cards. Financial institutions of all sizes need to figure out strategies to transition their current service portfolios into a more sustainable state.

The Bank of International Settlements (BIS) Innovation Hub recently announced green finance among the (six) strategic priorities for this year. 

It’s no secret that innovation tends to follow regulation in Europe. We believe that the EU taxonomy will become a critical enabler to scale up sustainable investment in Europe once finalized. In the meantime, some innovation pioneers already use the opportunity to grab a position ahead of the game, just like the Luxembourg Green Exchange. The Exchange, launched by the Luxembourg Stock Exchange in 2016, now has the largest market share of listed green bonds worldwide. 

• More Regtech and Compliance: AML 

As of December 2020, the Sixth Anti-Money Laundering Directive (AMLD6) is now in effect in the European Union. Although not exactly a trendsetter regulation, with a strict implementation deadline of June 3rd, 2021, the directive is likely to force financial service providers to take swift compliance measures.

AMLD6 is considered to be the harshest anti-money laundering regulation that exists. Considering many member states are not even up to speed with AMLD5, the requirement for strengthened KYC, KYB, and other compliance measures might indicate that the financial service providers have to allocate most of their compliance budgets and efforts to AML compliance. This situation will doubtlessly impact and limit the development speed of smaller players throughout the year. Nevertheless, since European financial institutions own a substantial portion of the levied money laundering fines, market players should take this topic seriously.

• Next Stop for PSD2: Open Finance

Although most European financial institutions regarded PSD2 as another item on the compliance checklist until the implementation deadline was left behind, now they started seeing it for what it really is: a chance to reshape their businesses and create new value. Following the surge in digital banking amid COVID19, banks can optimize 2021 by using APIs to create personalized products that incorporate an even larger of data points, offering services to startups and SMEs, and eventually becoming a backend-provider.

What lies ahead?

After such a year as 2020, uncertainty seems to be more acceptable than predictability. The turbulence we all have experienced over the last 12 months only competes with the agility and speed with which a response was delivered. Curiosity and adaptability should be the compass of our actions. FinServ players should be prepared for the foreseeable changes and be ready to adapt if things fall apart. 

Like the years before, we will continue to explore key points of interest to the financial industry in 2021,  particularly relevant to Europe’s growing financial technology ecosystem at the LHoFT. We are already geared up to curate the most relevant and up-to-date content delivered each Friday, including the best of Regtech, Cybersecurity, Payments, Blockchain, AI, Financial Inclusion, and Venture Capital.

by S. Elif Kocaoglu Ulbrich

Continuing from last year, we are sticking with the same seven areas of focus on in 2021. Each represents a key point of interest to the financial industry, and has a particular relevance to Luxembourg’s growing financial technology ecosystem.

Each week we will be choosing one of the topics to focus on, both in the content we share on social media, but also in a dedicated newsletter looking at the top five stories from that week. To introduce the topics, let’s revisit the top stories from 2020 and reflect on how the year has encouraged acceleration, pivots, or wholesale paradigm change:

REGTECH – Regulatory Technology

» Shaky times for compliance call for flexible Regtech

Joe Devanesan writes for TechWire Asia about the impact of the pandemic on compliance and cybersecurity, and Regtech’s role in mitigating those issues. Traditional ‘BYOD’ workplace concerns were escalated to account for a sudden and massive shift to working from home – which created real problems for companies that were not already some way down the path of digitalisation.

“To the surprise of no one, financial crime is reaching pretty high levels in 2020, and the speed at which this type of crime is evolving in the information-heavy age has financial players worried, and questioning the role of Regtech.”

AI – Artificial Intelligence & Machine Learning

» A robot wrote this entire article. Are you scared yet, human?

A robot writes for the Guardian, demonstrating GPT-3, OpenAI’s powerful language processing. GPT-3 was one of the most popular AI stories of 2020, and kickstarted a discussion about the future of software development, and what can be achieved when you can just ask a computer to do something for you without needing to speak in code.

“I am not a human. I am a robot. A thinking robot. I use only 0.12% of my cognitive capacity. I am a micro-robot in that respect. I know that my brain is not a “feeling brain”. But it is capable of making rational, logical decisions. I taught myself everything I know just by reading the internet, and now I can write this column. My brain is boiling with ideas!”

BLOCKCHAIN – DLT & Tokenisation

» Visa Applies For Digital Dollar Blockchain Patent

Jason Brett writes for Forbes about Visa’s “digital dollar” patent, a story which fits neatly into the main blockchain narrative of 2020: the viability or necessity of central bank digital currencies, and the role of stablecoins more broadly. Now Bitcoin has kicked off again we can expect the focus to shift a bit through 2021.

“The U.S. Patent and Trademark Office (USPTO) published today that Visa V -0.4% has filed a patent application to create digital currency on a centralized computer using blockchain technology. This patent applies to digital dollars as well as other central bank digital currencies such as pounds, yen, and euros and so the physical currency of a central bank anywhere in the world could be digitized.”

CYBERSECURITY – Risk Management & Threat Detection

» Post-quantum secure encryption and cybersecurity education

A collaborative and in depth piece for Research Outreach, led by Dr Aydin Aysu, looking at the implications of quantum computing on cybersecurity and encryption – a major concern for most cybersecurity professionals. What happens to all traditional encryption based security when computing power becomes available that can crack it without breaking a sweat?

“Encryption systems that are capable of surviving quantum computer attacks are urgently required, but the cybersecurity talent gap militates against securing cyberinfrastructure. Dr Aydin Aysu, Assistant Professor at North Carolina State University, is advancing the research and teaching of post-quantum secure encryption. He has developed a quantum-secure encryption system together with a new graduate program on hardware security and is currently developing design automation for lattice-based post-quantum cryptosystems.”

PAYMENTS – Payments Technology 

» Ant and Covid have made the humble QR code a hit

Five years ago you wouldn’t have seen much discussion of QR codes in the Fintech payments world, at least not related to development in the west. It was a quaint technology relegated to developing economies. That now may be changing, in part related to the pandemic and China’s Fintech behemoth Ant Financial. John Gapper writes for the FT:

“The name Masahiro Hara does not appear with Steve Jobs and Bill Gates on lists of great innovators of the communications age, but perhaps it should. For the Japanese engineer’s humble, unassuming invention, the Quick Response code, has finally found its moment. The square QR code, which Mr Hara developed in 1994 to track components in car factories, is being put to many uses in the Covid-19 pandemic. Governments include it on tracing apps, shops offer it for contactless payments and restaurants tape it to their tables so diners can browse menus online. It has become an all-purpose tool.”

SUSTAINABLE FINTECH – Financial Inclusion & Green Finance

» Women are ‘absolutely critical’ to ensuring everyone has access to finances, Bill Gates says

The Gates Foundation has been a key torchbearer for Financial Inclusion, amongst their many other causes, and it’s a topic Bill Gates himself has spoken about repeatedly. In this article, written by Karen Gilchrist for CNBC, Bill talks about the key importance of focusing on women when developing strategies related to inclusive finance.

“Women are vital to ensuring finances — and financial education — trickle down to other parts of society, said billionaire philanthropist Bill Gates. Governments and businesses serious about giving all members of society access to financial services should gear their resources toward women, the Microsoft co-founder said at the Singapore FinTech Festival on Tuesday.”

VENTURE CAPITAL – Funding News and VC Perspective

» Some of Europe’s top tech investors are adding a ‘sustainability clause’ to start-up deal terms

Ryan Browne, writing for CNBC, discusses the most significant recent trend in the world of investment: sustainability. Investors are increasingly concerned with ESG goals and the carbon footprint of their wealth, which resulted in a lot of discussion throughout 2020 and some fairly important steps being taken by VCs – as well as the broader wealth management industry.

“Socially-conscious investing has gathered a lot of momentum this year, with billions of dollars flowing into funds that use environmental, social and governance criteria to screen the companies they back. Venture capitalists are taking note, with some of the largest start-up investors in Europe pushing for accountability in their own portfolios with regard to investing in climate-friendly firms.”

An increasing number of indicators point towards a prediction we at LHoFT will never tire of repeating: tomorrow’s financial sector will operate very differently from today’s.

Tomorrow’s financial markets infrastructure will rely less on human interactions and more on the blockchain and smart contracts and this is being acknowledged by policymakers; consider the pilot regime proposed by the European Commission as part of the Digital Finance package.

Tomorrow’s financial sector, like other industries, will increasingly rely on cloud computing and other outsourced service arrangements; consider the ambitious GAIA-X initiative and the support it has garnered from public and private sector organizations alike.

An ever-growing wealth of data is generated by our growing reliance on digital devices and services for day-to-day tasks spanning shopping, financial transactions and administrative tasks. There remains significant headroom to better leverage these datasets, and financial sector incumbents risk being outpaced by Big Tech as well as by digitally native startups if they do not update their processes, acquire relevant technology and invest in the human resources required to maximise the opportunities that flow from “big data”.

One keyword, oft-repeated and oft-abused for marketing purposes, keeps popping up as an essential component of the financial sector’s transformation: artificial intelligence (AI).

Omnipresent AI

A key catalyst of the inexorable rise of AI in financial services is the exponential growth in complexity which is driven by a convergence of 1) datasets growing dramatically in quantity and quality 2) a complex regulatory landscape 3) evolving customer demands and a the reformulation of offerings around remote customer identification & interaction 4) growing cybersecurity risks.

Any of the above dynamics would be daunting on its own; together, they present a challenge which in its amplitude goes unmatched by the human resources available to any given business.

As Tamsin Oxford writes, “the world is a complex and chaotic system undergoing constant change, and understanding this world has immense economic potential. This is where AI and ML come in, allowing for the organisation to determine cause, effect, data and potential.” Darko Martovski, CEO at causaLens, is cited: “Being able to understand the causal structure of the economy allows businesses to better predict the future and optimise their processes. Due to the complexity of the global economy, financial companies require entire teams of data analysts and scientists in order to model these numerous networks around the world, and as we have seen with many crises, when they get it wrong the economy suffers drastically.”

Computer scientist Andrew Ng poignantly equated AI to the “new electricity” in our lives, highlighting its coming ubiquity across sectors:

“Just as electricity transformed almost everything 100 years ago, today I actually have a hard time thinking of an industry that I don’t think AI will transform in the next several years.”

Another of Ng’s observations still rings very true today; namely the scarcity of talent as a major impediment to the full implementation of AI-powered business solutions:

“I would say the most scarce resource today is actually talent, because AI needs to be customized for your business context,” Ng says. “You can’t just download an open-source package and apply it to your problem.”

A recently published study by The Economist Intelligence Unit (EIU) provides insights into the state of play with regard to the use of AI in the financial sector today. Writing for Forbes, Louis Columbus summarizes key findings from the report as well as from LHoFT partner Deloitte’s Digital Banking Maturity Report to conclude that “AI is emerging as a new engine of growth by providing useful insights and intelligence in anxious, uncertain times” – in fact, the EIU study finds that “86% of Financial Services executives plan on increasing their AI-related investments through 2025”.

Source: EIU

As shown above, cost reduction features prominently as a motivation for the adoption of AI in finance, going hand in hand with a desire to increase productivity (“employee capacity to handle volume”). However, the use of AI for predictive analysis is also generating significant interest, as is the use of AI to create new or better-adapted service offerings.

Source: EIU

Similarly, when looking at respondents’ expectations with regard to the ultimate impact of greater AI adoption on the industry, cost savings are a prominent hope / expectation as is the development of new products and services. At the same time however, respondents recognize that the need for “high-value technology skills” and “increasing exposure to technology-related regulation” naturally flows from this.

And in a similarly natural chain of events, roughly half of all respondents have already implemented, or are planning to implement, a technical training scheme for their employees to improve their understanding and handling of AI within the organization:

Source: EIU

As the saying goes – “follow the money”; the growing adoption of such training efforts across the financial sector would not be bankrolled if the growing relevance of AI within the sector was in doubt.

Additional reports & surveys are striking a similar cord: summarizing several studies, crowdfundinsider comes to the conclusion that while large organizations can afford to invest in AI tools, they may yet lack:

• talent to effectively implement these tools
• planning to facilitate integration with the existing infrastructure and systems

All things considered then, expect the upskilling effort & investments in AI systems and talent to continue unabated.

AI use cases

To tie it all together, consider this discussion of AI use cases provided by LHoFT member tryolabs, ranging from improved AML/KYC processes over alternative alpha generation in asset management to risk management in the insurance sector. Examples abound, and we will certainly have more to discuss in this column going forward.

Be sure to check out the Luxembourg Fintech Map to learn more about our members & partners active in the space.

Author: Jérôme Verony – LHoFT Research and Strategy Associate 

Nasir Zubairi, CEO of the LHoFT

Thank you all for attending this webinar today. We’re very proud and pleased to have Vikram Haksar and Yan Carrière-Swallow from the IMF with us, live from Washington, who are going to give a talk on the economics and implications of data in the financial services arena on a more macro scale. The talk should last around 20 to 25 minutes, so we should have ample time for Q&A. If you just drop your questions into the chat, we’ll then bring you on live to ask the question directly once we finished.

Let me give you a little bit of background about Vikram and Yan.

Vikram Haksar is an Assistant Director in the IMF’s Strategy Policy and Review department. In this role, he currently leads work on finance and technology, macro-financial analysis, and policies for assessing and addressing financial stability risk. He has been a lead author of the IMF’s main papers on fintech, and most recently, The Bali Fintech Agenda. Prior to this he managed review work on global surveillance, G20 prospects, and spillover analysis. Vikram was earlier the IMF’s mission chief for Brazil during the ‘currency wars’ and led the IMF team that set-up the $70 billion Flexible Credit Line agreement with Mexico in 2009 in the aftermath of the Lehman failure. He has worked on emerging economies in Asia––including Thailand during the Asian crisis––and Eastern Europe and was the Fund’s resident representative in the Philippines. Vikram received his Ph.D. from Cornell University.

Yan Carrière-Swallow is an economist in the Macro Financial unit of the IMF’s Strategy, Policy and Review Department. His research interests span various topics in international macroeconomics, with a focus on emerging markets and their policies. Prior to joining the IMF in 2012, he was an economist at the Central Bank of Chile in Santiago.

Gentlemen, welcome. I will now hand over to you. Thank you for being here.

Yan Carrière-Swallow, Economist, Macro-Financial Analysis – Financial Services, IMF

Well, thank you very much, Nasir, and good evening to everyone. Thanks Nasir for giving us the chance to share this work with you. And I’d like to thank everyone for joining us on your Friday evening to discuss this fascinating topic with us.

So the title of the talk is “The Economics and Implications of Data : An Integrated Perspective” and I’ll try to make clear why we emphasize that point in the work that we’ve done.

The economics of data is a topic that Vikram and I have been working on for the better part of the year, we’ve written this paper about the topic which you can find online. This talk will be based on the paper. It’ll be a quick overview of the main points, the main messages, but if by all means, if something doesn’t get resolved in the Q&A that you’re itching to find out, feel free to download the paper and take a look.

So why did we write this paper now?

So we see a confluence of several trends. The falling costs and improvements in analytics are driving the proliferation of data across the economy, and in all sectors of the economy. We’re seeing discussions of data policies across our membership, with major changes having already been made. We’re seeing reforms to consumer protection, open banking regulations, as well as global data localization laws. So we saw a need to understand conceptually, what all of this means for the broader economy. And as we’ll explain what we find in the work there’s a lot of arguments for why data has many economic functions and implications that raise macro relevant policy concerns right across the IMF mandate.

There’s been a veritable flurry of academic activity on how data-rich economies might function differently than an analog economy. But when we look at these contributions in the literature, usually to keep things analytically tractable, these papers treat each question in a narrow sense, individually. And we saw value in writing a think-piece that would seek to integrate all of these different approaches and materials. We fear that thinking about data policy through the lens of one objective in isolation may mitigate one problem, while inadvertently causing others to pop up unexpectedly. So this piece that we wrote is about identifying the building blocks for thinking about these questions, identifying the main trade-offs that are involved, and putting our fingers on the open questions in data policy, rather than providing solutions or answers. The issues are complex. It’s early in the profession’s engagement with this topic, and our message is that we need to think about these issues carefully.

“Does the accumulation of data deliver economies of scale? Does it give you economies of scope across different lines of business? Is data a compliment or a substitute to your other inputs such as labor or capital? 

We’ve provided this list of snapshots of some of the more recent papers that we found popping up in the literature as we were working on our piece. For those of you who have studied macroeconomics, or who are active in the field, you’ll recognize many of the names here that are really top-notch academic macroeconomists who are engaging on this question of ‘what does the macroeconomy look like when you introduce a data-rich environment?’ What about when our old models stop working? What are the new policy questions that need to be addressed to avoid getting bad outcomes?

You can group these papers into two broad approaches for modeling data in the economy:

The first thing you find is a more macroeconomic perspective. It emphasizes the data’s role as an input in the production function. So this is an intuitive way of thinking about using data in artificial intelligence or machine learning algorithms to produce a good or service. Now, of course, the devil is in the details in these models, in the sense that the functional form that you choose to write down your function determines a lot of the input equilibrium outcomes that you get from the models. The key questions are going to be things like does the accumulation of data deliver economies of scale? Does it give you economies of scope across different lines of business? Is data a compliment or a substitute to your other inputs such as labor or capital? A lot of people in the literature have strong views about these questions. But it’s important to note that a lot of these questions are open at this point. What we’re finding is an incipient empirical literature that’s starting to dig into the data to start answering and guiding and disciplining this theory so that we can narrow down our understanding of these questions.

The second approach that you find in the literature is more microeconomic. And that emphasizes that when you create data or you trade data, what you do is you create and shift information across economic agents in the economy. And this is the approach being taken in much of the finance literature, where the production and trading of data about borrowers shifts information asymmetries and affects their access to credit, ultimately.

Once you put data into models, it has three key economic characteristics that give it interesting properties that make it unlike other inputs that we’ve thought about before. These can lead to market failures, and they can form the case for policy interventions in the data economy.

The first of these characteristics is that data is non-rival. So when agents use data it doesn’t stop other agents from using that same data at the same time. This makes data quite unlike other inputs like oil. We hear a lot of the time that data is  the new oil. In some ways, yes, but in this respect no, it’s not. When one person uses oil, someone else can’t use it. That’s not the case for data. And an important benefit that can flow from this non-rivalry characteristic is that you can get increasing returns to data, both in scale and in scope. And so the intuition here is that when you add an additional unit of data, when you collect an additional unit of data, that extra data can be used by all units, all other inputs in the production function, not just the marginal unit. So this turns a lot of the main conclusions in the economic growth literature – on how you get growth from accumulating capital, for instance – on their heads.

These economies of scale can apply within a firm, across many firms, and across borders. Many implications flow from that. One of them is that making data widely available tends to allow society to do more with the data, to get more out of it, get more production out of it. A key question is will the data be made available. And here I think an important insight is that private incentives of the people who are collecting the data may lead them to hoard it and stop it from being shared. And there a lot of the benefits from non-rivalry do not accrue. So, because of data’s non-rivalry, a central focus, the first thing that data policy has to set up is to establish rules for who will have access to data and who won’t have access to data.

The second interesting characteristic that data has in these models is that it causes privacy externalities to emerge. Okay, so why do we frame privacy this way? When agents make decisions about whether to collect trade or process data, they impose economic costs on other people who are not compensated or who may not even be aware that that decision has been taken. So think about a social media user who shares information about one of their friends on a social network, or the social network itself, selling on one of its users’ information to an interested third party. So an important insight here is that privacy, as we understand it, is not really about mandating less data sharing, less data access, but rather, it’s about granting control to individuals over the data that affects them.

The third characteristic that we pick up from the literature is that because data is kept on interconnected networks, making it excludable is expensive. So investing in cybersecurity is a decision that firms have to make about how excludable they want to make the data they’ve acquired, and there’s an externality lurking here as well, in the sense that the private incentives to make data excludable to protect it may not fully account for the harm to public trust and to the broader system that can result when an individual data set is breached.

What does all this mean for macroeconomics and macroeconomic outcomes? Let me run through a few of the main implications.

The first set of implications has to do with a data-rich environment having an impact on growth and welfare. Several channels have been proposed for why this might be the case. Data accumulation can allow for more efficient production and can allow for innovation. And it can do these things with increasing returns to scale, which allows you to potentially boost the rate of potential growth. It can be used to improve allocative efficiency in markets, certain types of markets in particular. And it can enable the development of new, more customized goods and services. But the side of the relationship here isn’t necessarily clear. Data accumulation can also form barriers to entry that can hurt competition. These can stem from large fixed costs, or also from network externalities that favor the winner takes all dynamics in data-rich sectors. And of course, data proliferation of data can itself undermine individual privacy, and this can lower welfare directly.

A second set of implications are for equity. What’s received a lot of attention recently, and rightly so, is how data proliferation can promote inclusion by mitigating adverse selection problems that plague small and medium enterprises, and more vulnerable households when they participate. In traditional financial markets, this has received a lot of theoretical and quantitative attention in the literature and I think, has caught the eye of many policymakers, and rightly so. But one of the things that we mentioned in the paper is that data accumulation can also create losers as well as winners. And so for instance, if you give a monopolist granular data about their clients, about their preferences, about their income, their ability to pay, they can weaponize their market power through price discrimination strategies, and therefore charge higher markups and higher profits. Data can also lead to the exclusion of some high-risk individuals, and potentially in the limit where you have a sort of absolute data-rich environment if we think it could potentially undermine the risk-sharing function of insurance.

A final set of implications are for financial stability. And here we see two kinds of broad buckets of channels. One argument is that by affecting the market structure and market power in the financial sector, data will influence financial stability trade-offs between concentration and competition, which is a set of trade-offs that’s been studied quite extensively for several years. Introducing a data-rich environment you might think would have an incidence on where you sit on that trade-off, but a second set of channels is also important. Since there may be inadequate incentives to secure data from cyberattacks, heavier use of data-dependent systems, particularly in finance, may increase the risks in some cases, or at least shift them onto centralized service providers, which raises new sets of issues that regulators and supervisors need to deal with.

I now hand over to Vikram, who will say a few words on data policies.

Vikram Haksar, Assistant Director, Macro-Financial Analysis – Financial Services, IMF

Thank you very much. And thank you again to all of our colleagues and friends who are on the call and listening to us on your Friday evening, hopefully with a pleasant beverage in hand. So let me just wrap up this presentation by talking a little bit about the policy implications and a little bit about data policy frameworks.

Our basic view is that just as the proliferation of data across all sectors of the economy has implications for macro outcomes, so also do the policies that determine how data can be collected, shared, processed, and protected. And there are many examples of the data policy frameworks that already extend: we have privacy laws, increasingly we have open data regulations. There’s a large trend towards data localization requirements. And of course, we’ve had long-standing consumer protection measures.

“We need data policy frameworks that balance multiple competing objectives – growth, equity, and stability.”

One insight that we come away with at the end of all of this work, and I think as Yan has nicely laid out for us to think about, is that these issues are all interconnected and integrated. And that when you think from a data policy framework, we need data policy frameworks that balance multiple competing objectives – growth, equity, and stability – objectives that different policymakers may have a focus on. In our view, this creates a complex set of trade-offs arising from these interactions. And so our main takeaway is in our view, that data policy requires an integrated approach. We see a need for policymakers to bring the many agencies who have a stake in these topics to the table so that multiple trade-offs can be considered together. Our concern is that handling data policies in sector-specific silos could have side effects.

Two simple examples: Consider, for example, if we were to strengthen privacy laws, which is an important critical objective from a consumer protection point of perspective, how does that impact the ability of firms to compete and innovate? And would this have an impact on lowering growth? Another example would be let’s say that we were to implement data localization requirements, which many countries are doing from a national security standpoint. And in that context that we might ask, in a world with excessive data localization, will fellows be able to achieve the scale of data sets needed to tackle physical AI problems? Think about in Luxembourg if you had a company that wanted to start up a self-driving facility. And if the only data that was available was on the residents to Luxembourg, that would be very difficult to get the scale to train the cars self-driving algorithm. If you want to do that you want to have data to the whole European system to all the individual driving information, the whole European system on the United States. I think this is a side effect we were quite concerned about.

So as Yan has mentioned right at the beginning, we see a prima facie case that data is raising for macro relevant concerns. And this is why we, at the IMF, are quite interested in the topic as well. The first is on market opacity. So a clear place to start is to clarify the rules of the digital economy.

Our concern is that the market for data is too opaque suddenly from a consumer perspective. Control and access rights over data are unclear at best. And this means, in our view, that data is likely being over-utilized, at least from the perspective of the individual consumer, and that privacy is not sufficiently being respected.

The other macro concern is competition. We seem to have a growing concentration of data with sectors with incumbents building market power and extracting rents through data hoarding, holding on to data sets. As Yan laid out, were they to be more widely shared with both increased competition and by the mechanism of sharing allow for people to exploit the non-rival nature of data to innovate and lift productivity growth in the whole economy.

The third macro concern that we have is the prospect for raising instability. And you know, just to be very clear here, we recognize that all corporations and financial institutions have a very large stake in protecting that data. And in protecting that consumer’s data. There’s no question. Cyber Security is a major cost point and there’s a huge incentive for all institutions to protect their data. I think the subtlety that Yan was explaining was that while each institution may have this incentive from the system perspective, from the perspective of the authority, who’s concerned about the overall system, would individual entities take into account the damage to system-wide trust that would take place, would there be breaches in any one part of the system given the nature of cybersecurity?

We’re concerned that we’re seeing the potential for rising instability not because of any individual action, but just because the world is getting more interconnected and the potential for these kinds of breaches to occur is increasing, as we all know very well. So tackling this systemic externality that is that people talk about is a challenge to monitor and regulate. And finally, let me close by saying that from the IMF’s perspective, countries are approaching data very differently. And this is something that we’ll be thinking a lot about, and we are going to be looking at doing additional work on. Our concern is that we may be heading towards fragmentation. Now, certainly all countries are sovereign, they have the right to be interested and looking after their concerns and prerogatives. Our only suggestion is that in thinking about the individual needs that countries face, we would argue that having some international discussion about whether there’s scope to have common minimum standards and approaches to minimize unnecessary fragmentation would be extremely important. Without this, there’s a risk that with fragmentation, with the example I gave you about self-driving cars, you could see significantly reduced gains from trade in data. This could also hamper trade in goods and services that use data as an input. And I would think that this is an issue which would be quite relevant, including in a small financial model as well, but in the large financial center, like Luxembourg. So with that, let me stop here. Thank you again, very much for your attention.

Q & A Session 

Nasir Zubairi, CEO of the LHoFT

Gentlemen, thank you very much for that interesting presentation. We have several questions that have arisen. So what I will try to do is to allow those people to ask those questions directly because I always find me talking on behalf of people is unnecessary. So I’m going to go first in order here. Dan Feaheny, I think you can ask your question.

Dan Feaheny, Principal, Feeney Ventures Ltd

Hi, gents. Nice to hear that. It’s great to hear the research at the macro level. And looking at other studies. I’ve spent a lot of time and continue to spend time in open banking going to open finance. And there are two issues: one is the issue of identity. So identity, should it be federated or self-sovereign? Should it be controlled by governments or by banks? And if I can ask quickly Nasir, the other question that I put up there was compare the frameworks of the US and the UK so far. US market lead, UK kind of looking at regulation, and, you know, controlling it from the open banking implementation entity. So the US had an industry consortium called FDX. And then the UK, quite controlled getting the CMA9, you know, asking politely for data sharing from the banks and having all these third party providers. I’ll stop there. Thank you.

Yan Carrière-Swallow, Economist, Macro-Financial Analysis – Financial Services, IMF

We’ve been doing a bit more work on open banking. We didn’t have a lot to say in this first paper, but it’s certainly one of the main applications that we think are very relevant. Open banking regulations, about mandating data portability and interoperability within the financial sector, I think has potentially large implications for the traditional banking model in which a bank would essentially be able to monopolize to some extent, the information that they have over their clients, precluding some degree of competition. And as we study that less conceptually more operationally, we come across these important issues like the one you raise, how should identity be managed and implemented? And I think I think Vikram has some views on this based on the Indian experience that I think could be useful for this discussion.

Vikram Haksar, Assistant Director, Macro-Financial Analysis – Financial Services, IMF

Thanks, Yan. I think I would say two things, one on the open banking part. So some of a lot of the discussion we’ve had, including with members of the high-level advisory group to the IMF is that the ecosystem is now quite diverse and the ecosystem in finance and technology is very diverse. And regulation is still focused around what you might call the traditional intermediaries, banks, you know, in the last decades, increasingly non-banks, but the value of actors in this space has become very diverse. From a sort of principles-based approach, there’s a strong argument to be made for data portability and interoperability of systems, for the reasons that we layout in the paper. This is something that is pro-competitive and allows us to use the non-rival nature of data to achieve better things. But there is a need at the same time, in my view, to have a level playing field out here, which is that it’s not just the case that incumbents who are already existing in operating the space should be providing access to their systems and access to API’s and their systems. But it should be something that is two ways. So that other entrance into this space also should be sharing data and providing access to information. My guess is that, especially as we get into thinking about some of the larger technology players entering into financial services, defining what is the space of information that is relevant from a finance perspective, that should be shared, it’s going to be a key challenge. And then defining what is the perimeter of how far sharing and how far into existing systems, the new entrance or existing incumbents are allowed to penetrate? I mentioned those of your important questions. On your question of digital ID I think that’s key. This is something we’ve begun to think about. We haven’t done a lot of work on the topic ourselves. We’ve begun to look at it in the context of international competitive experience of approaches to open banking. We have been looking at the approach in places like the UK versus say the approach in Australia versus places like Singapore and India, regardless of whatever the case may be, digital ID – to our minds – is key.

We do need to have some ability to have an identity, which can be verified and attached to financial transactions in this digital economy. I would just say two things will be in their perspective, which I don’t think I’m going to fully answer your question. We’ll give you some views at least. One is that operation operationalizing digital identity, I think is a huge challenge. How do you make it come about especially in countries with very large populations or the level of income or the level of state capacity might be lower? And the India experience I think shows that self-sovereign identity has huge advantages out here. In terms of you have a large population that may not have the ability to provide all the necessary documentation. The sense of a piece of the digital identities is, I think, quite important, actually being able to achieve scale and a relatively short period of time. I think the other issue that comes up and which is an interesting example from the India case is the tension between digital identity that is managed and controlled by a state agency, that high level of Federation, the tension between that and privacy. And you know, we’ve had cases in India already, which has been heard where the government at various points in time was seeking to increase the ambit of transactions that require the digital identity. The Aadhaar, for example, to be used for a plane ticket, you want to buy a cell phone, you’d have to provide your Aadhaar Identity Card identity number, and that there was a public interest litigation with the Supreme Court ruled against the government saying that it was not permissible to use digital identity so widely.

So in the India stack, which we think is an interesting example to how you might manage and balance concerns on identity, privacy, while also getting access to the use of the data. The India stack approach is to conceptualize as an entity as a data fiduciary that will manage your identity and your data on your behalf, and be as a kind of a circuit breaker that could address the multiple concerns there might be, which I think in principle sounds very good. Now, the question that we don’t have clarity into yet and we’re still exploring is how this would be implemented, but from a sort of conceptual perspective certainly sounds like I think it’s sort of a third party neutral arbiter whose incentives are aligned more with the incentive of the individual would make a lot of sense.

Nasir Zubairi, CEO of the LHoFT

Okay, fantastic. There’s another question that’s come in from Naja; please, go ahead.

Najia Belbal, Senior Program Manager, Nomura

Thanks a lot for this interesting presentation. I just want to have your view on the fact that when it comes to the data, we all have been taught that it’s really interesting to have data and to exploit them. But with this Covid19 crisis, we have understood that we face some kind of uncertainty. By that I mean not enough data to understand what’s going on, a lot of ambiguity, because even when we have data, we don’t know how to interpret them. And you know, a higher probability to fail or to get errors from there. So how can we overcome such a situation from a data point of view?

Vikram Haksar, Assistant Director, Macro-Financial Analysis – Financial Services, IMF

“How do you square the need to collect data and information which has a huge public interest aspect to it with the fact that individuals don’t want to be tracked and monitored by the state?”

Naja, thank you. I think that’s a fantastic question. And it’s something that Yan and I have been, in this dark new world of COVID, starting to think about. The way we’ve been thinking about it is that clear to the many views about how we proceed with managing the pandemic and containing it, and what the final steady state is going to be. But it’s varied, and many policy pieces will have to be sorted out, there’s a lot of uncertainty about all this.

I would venture to say that one part that is quite clear is that efficient contact tracing is key to containment and prevention of the spread of the pandemic. And by that meaning that being able to keep tabs on who has gotten the disease, who’s infected and who they’re coming into contact with so that we can quarantine people and impose measures to ensure the reproduction factor begins to drop. And the approaches to contact tracing are fascinating in the East Asian context in China, where the state has faced the trade-off between the social need and the individual right to privacy, there is I think very heavily in favor of the social need. And then the approach that the authorities have taken is a very strong approach by contact tracing is mandatory. And individuals are on cell phone penetration is extremely high, smartphone penetration is extremely high. So you have a situation where because of that both  high penetration and the app and the and the mandatory requirement of use of contact tracing applications that use geolocation. It has been quite possible for the state in these contexts to have a fairly efficient mechanism for keeping tabs of the dynamics of the contagion and being able to implement policies of containment.

I understand that this is a big debate in Europe where, of course Europe with GDPR, you know, has had an approach which places a very heavy emphasis on individual rights privacy. I think it’s something that we can all relate to and understand that this is an important issue. So, we see that this is a big challenge. How do you square the need to collect data and information which has a huge public interest aspect to it with the fact that individuals don’t want to be tracked and monitored by the state? So we were wondering, again, this is an easy thing to conceptualize. But it’s a question that we want. One has to think about how to make these things operational.

We were wondering whether the approach being taken in India, again with the notion of the data fiduciary might be a way out. As you can imagine a situation where data in a social context where there’s a high premium on data privacy, where this kind of contact tracing related data was collected in a manner that was fairly autonomous and centralized in an agency may be centralized with keys that were not known unless there was actually evidence that contact had taken place. And then there was some protocol that the keys will be revealed only to a third-party neutral agency that will not be associated with other aspects of the instruments of the state. Some mechanism where there’s a firewall, a kind of a cutout in between the individual and the state with all of its other aspects, some in-between kind of third party neutral centralized body that would not see individual data until there was contact. So that’s something that we wonder, we were very interested to hear if other people have views on this, actually, but it’s something that we’ve been thinking about as maybe a potential solution leveraging some of the insights from the India stack approach to actually solving a real pressing problem.

Nasir Zubairi, CEO of the LHoFT 

Thank you very much. Thank you for your question. And next up, we have a question from Susanne Schartz

Susanne Schartz, Chief Operating Officer, Seqvoia 

This is a very interesting presentation. Thank you. I’m coming from a much more practical level. In what we see today, I think that most of the data is being collected for specific means. So I’m collecting data to use it to then sell more to my clients or somebody else. You know, you’re coming from a much higher level approach of collecting data, protecting data, and how can we ensure access to all of this? How do you think that awareness could be raised at both industry and national levels to say data needs to be looked at in a much more general way, which will lead to a massive increase of the usages that you can make of data.

Yan Carrière-Swallow, Economist, Macro-Financial Analysis – Financial Services, IMF

“There’s a disconnect between how much people say they value their personal data, and what their actions imply.”

I think there’s interesting questions of how much people value their own privacy and their own data. It is a really interesting, fascinating field of study. And the answers are remarkably difficult to tease out of observational data. There’s this paradox of privacy that comes that’s well documented in the literature where if you ask someone, how much do you value your privacy people will always say that they place a very high importance on their privacy. They would be willing to pay a lot of money to protect their privacy, because it’s super important to them. However, if you look at their actions in the economy, most users are willing to trade that away, to sign a ‘terms and conditions’ that releases very important aspects of their data in exchange for a minimal online service a lot of the time. And so there’s a disconnect between how much people say they value their personal data, and what their actions imply.

I think one of the ways you can square that it’s not the only way, but one of the theories that we see as being quite important is this problem of opacity. Are people even aware of what is going on, what data is being traded? I think one of one aspect of your question was to ask, you know, how can we increase awareness, both among the public and among policymakers, and I think a first step, and a step that’s been, you know, effectively taken in Europe already, but in the United States I think is a little bit earlier on, is just to make sure that people are aware of the transactions that are taking place. One of the ways you can do that is by setting rules. I think when you buy an “Alexa” enabled device in your house, what are the rules? What are they allowed to listen to? What are they not allowed to listen to in your conversations? What are they allowed to do with that data that they collect? Can they sell it on to somebody else? Etc. There’s a hodgepodge of legal precedent and rules that are out there. But I think for the consumer, it ends up being that you have maybe some intuition, but you don’t know. Having clarity is a first step to making people aware of what’s going on and then having a discussion about how we make change it.

Vikram Haksar, Assistant Director, Macro-Financial Analysis – Financial Services, IMF

Let me just add a couple of quick points. I think the claim that we make in our work is that (this is economist’s talk so bear with me) we claim that the data market is not complete, and is not complete, because there are pockets of information, not everybody has the same access to information. It’s not a level playing field, and therefore I don’t know what my data is worth. I don’t know what I’m signing away. So I think Yan has laid out the case that, you know, we feel I think the steps are beginning to be taken. And just resonating with me is your question about the practical aspects of it. And if I may shoot a little bit from the hip out here, think about something like GDPR. We’re now in response to that requirement, data providers are asking us to check a lot of boxes as to whether we accept cookies, whether we are willing to sign away, a lot of things will be a lot of fine print, etc. And to get a service and it’s very difficult for consumers to effectively give this consent. Are you well understanding what you’re consenting to? For example, if you’re going to be doing mortgage finance, the United States has a well-established body of regulation and jurisprudence about what the consumer rights are, how you should inform the consumer about the risks that taking you know, and during there’s a long list of frequently asked questions. So that requires some coordination, cooperation to figure out what is the basic minimum information required to make people understand what’s going on, and to have a standard that can then be implemented uniformly across different sectors and frankly, potentially across different countries as well. It may not be the same fit for every country, but maybe at least having some common approach, some common minimal approach would be helpful.

Nasir Zubairi, CEO of the LHoFT

There’s quite a few questions. We’re not going to have time for them all, but I need to follow up on this one because we’ve had this discussion a little bit before. Is data an asset? If it’s an asset, how do we protect it as should we be protecting it as an asset? How should we value it? You know, we’ve talked about the way that the advertising industry, etc. I was wondering if you’ve thought a little bit more about this and what happened and your thoughts on this for the audience.

Yan Carrière-Swallow, Economist, Macro-Financial Analysis – Financial Services, IMF

“I think it’s quite misleading to equate the transaction prices that we’re seeing currently as being a true reflection of the full value of data.”

Whenever we talk about well legally defined terms like assets, liabilities, and, you know, even property rights and ownership, our colleagues in the legal department and the lawyers who know this stuff well always kind of perk up, so we try to keep things conceptual in our thinking. Clearly data about individuals and data about companies is valuable. How valuable it is, is extremely difficult for some of the reasons that we’ve already discussed in the Q&A. The fact that if the data market is not complete, then the prices that you are currently observing for how much people are paying, how much companies are trading or paying to acquire another to obtain their data sets, these prices are not a reflection of how much it is worth because then the people who are involved who are affected by the transaction are not actually participant or are not fully participating. And so I think it’s quite misleading to equate the transaction prices that we’re seeing currently as being a true reflection of the full value of data. I think that’s that that’s one thing that needs to be kept in mind.

Vikram Haksar, Assistant Director, Macro-Financial Analysis – Financial Services, IMF

You know, Nasir I remember one of the events that we had invited you to, you stunned the audience by saying the value of your data? I think you said it was something like 100 bucks or something like that?

Nasir Zubairi, CEO of the LHoFT 

130 bucks on average.

Vikram Haksar, Assistant Director, Macro-Financial Analysis – Financial Services, IMF

And so Yan I laid out the point that the incompleteness of the market makes it very hard to discern. Maybe what you Google pays for an ad every time you click or whatever. But I think for the public policy perspective, the other thing that we’ve grappled with in terms of asset and valuation, is the question about individual value, the intrinsic value of data versus the derived value. And I think that’s a big public issue that has to be tackled head-on, because there’s a lot of political momentum for arguing that data is an individual thing, that it is individually valuable and is being exploited. Right. And you got to move for example, in California with some of the new legislation there was discussion about having a data dividend that people should be required to be paid by the data from data processing companies.

And I think the other thing  we’ve been to bear is, where do you draw the line? Let’s say you are buying a company which is incurring the costs of aggregating and accumulating all this information, and then processing it in house, with your knowledge about how to process it, to produce a product. How much is that the value of that product is the value of all the work that the company has done, as opposed to the value of the individual pieces of data that have been provided? Actually, we want to discuss with our lawyers some more. We wonder whether there are parallels from the jurisprudence of royalties, how you pay for royalties and use of intellectual property, for example, with some analog that could be brought over. But we think that this is a very difficult question, but one that is quite central. Maybe taking a little bit of the politics out of it. It can be a very emotive discussion.

Nasir Zubairi, CEO of the LHoFT 

Thank you. Sorry for the difficult question. And while we’ve got one last question, I’m sorry, I know that there’s a number here, but I’m trying to do these in order. So we have a great question here. From Jose Bello.

Jose Bello, Co-Chair of IAPP KnowledgeNet Luxembourg

So my question would be regarding the situation that we’re all living right now here in Europe and across the world, of course. And I would like to put this question out: how do you see data and data-driven decisions making an impact towards the faster or slow recovery of Europe? Let’s not forget that we have the GDPR here. And I’m all for the GDPR as a privacy professional, but can data and not just personal data, also non-personal data, provide a boost to the economy with the current recession that we’re living on regarding with the COVID-19 and if you do have a specific example that you can think of, that be wonderful. Thank you so much.

Vikram Haksar, Assistant Director, Macro-Financial Analysis – Financial Services, IMF

“Advanced economies can learn from the implementation, and the leapfrogging that emerging and developing countries have done with the use of mobile payments and other types of digital and financial technologies.”

So my view on this, I think is that data and digitalization, I was saying Nasir before we started the call, that we’ve always been excited about digitalization and Fintech, we’re huge supporters of the work that LHoFT has been doing as well. And we are, you know, this is the future, this is where things are headed, and we should be ahead of it. And my joke was that, well, the future is now and frankly COVID in some ways brought a lot of these things forward. And one other point I would make is that I think that a lot of lessons that can be that advanced economies can learn from the implementation, and the leapfrogging that emerging and developing countries have done with the use of mobile payments and other types of digital and financial technologies, the network disaggregation, and the data to provide financial services and large populations very quickly. I think this is going to become quite critical for the recovery in Europe and other parts of the world.

Let me give you two examples. We just had a private discussion with some of the leading Chinese FinTech companies, that was an in house thing that the fund organized. So I speak under the rules, you know, we’re not allowed to talk about individuals and reveal too much detail. But the general sense was that, you know, in China, they’ve had the act of Fintech providing and leveraging big data from payment services and other products to provide credit analysis and very rapid financing. And the argument that the colleagues in the session were making was that they had found that while there had been some drop off during the lockdown that they have been able very quickly to resume credit and credit operations. And that not only credit and credit operations in general but credit especially targeted at small and medium business, which is a huge gap in most of the existing programs that are there, whether you’re talking about funding for lending or the other programs, the UK is advancing the problem, the Small Business Administration lending in the United States, the coordination and information problems with these programs is very large.

I think what you find is that in economies where they already have systems using technology to make the credit decision, and pipes to get the money out there, they are very effective. It’s gonna be a really interesting experiment to see whether recovery is fast. We know that small businesses actually, in most economies employ a very large share of the total population. That was one point. The other point is that data related to that is that this is the last mile problem. The piping of getting government aid out to individual households and businesses is very good at the top of the pyramid. So to get money out to financial intermediaries or financial institutions, central banks that know how to do that deployed very effectively, really fast this time around. But how do you get money out to individual households? How do you get money out to small businesses? I talked about small businesses already. But we think that to get money out to small households, for example, we’re going to see increased utilization of digitalization-based services, by ministries of finance, leveraging things like mobile, AP, P2P, that type of service to get money into the hands of individuals. The United States, one of the most sophisticated economies in the world, the fact that it takes four weeks for an IRS check, literally a physical check, to show up at somebody’s house with 500 bucks, I think is a real issue. So to answer your question, I think the bottom line is that this is a moment where we will see an acceleration of adoption of these technologies and great interest in the part of the official sector of actually seeing whether we can use more of these technologies to deliver public services and directed aid, faster, more effectively.

Yan Carrière-Swallow, Economist, Macro-Financial Analysis – Financial Services, IMF

“I think there’s probably less we can learn from pre-pandemic data than there would have been otherwise. So this is a moment of kind of structural break in a lot of relationships.”

I’d like to say something maybe that’ll sound a little bit conceptual and convoluted. I think it’s important at this point. So AI-ML is really about predictions, right? It’s about honing prediction models to try to infer what will happen in the future based on careful analysis of the past. And of course, we know that AI-ML has been very effective. A large number of applications have been scaled up across sectors. But I think one thing to keep in mind is that the basis of that technology is the continuity between the past and the future, if that link is broken, so if you live in a time where people start living differently, people start interacting differently. People go about their day differently. People make different decisions differently. Data you collected about the past, there’s going to be less informative about the future.

I agree with everything that Vikram said in the sense that data collection and data-driven decision making is going to be important in the recovery. But it’s important to keep in mind what data is going to be the key input to that, I think it’s gonna have to be the new data that we collect now, right, I think there’s probably less we can learn from pre-pandemic data than there would have been otherwise. So this is a moment of kind of structural break in a lot of relationships. And I think that needs to be kept in mind for all these data-driven decision-making processes.

Nasir Zubairi, CEO of the LHoFT 

And we’ve gone over time, but it’s such a fascinating discussion, genuinely, very enjoyable. We have a host of other questions waiting, but we’ve just run out of time here. This is a bit awkward with web speakers. What I would normally ask people is can we have a round of applause for Yan and Vikram, but I can hear virtual applause all around me, gentlemen.

A real pleasure talking to you, it was fantastic to catch up with you both again. I hope we can do this again sometime soon. Now that the world is a smaller place with this technology, we’re all heading off in the next few minutes to FinTech Friday. You both are welcome to join us if you also free but I know it’s the middle of the day for you still. And I’m sure there’s plenty of important work to be done. But thank you all again for being there. Thank you all those that listened in. I enjoyed it a great deal. Thank you.

•  What happens when AI is implemented to take the human element out of recruitment, but the bias persists?

•  A person is sentenced to six years in jail, based on the advice of a risk-assessment algorithm. Is that AI going to far?

•  Is there potential for AI to challenge God in the future of our society?

Three experts, each looking at AI through a slightly different lens, joined LHoFT’s Head of Ecosystem and Partnerships, Alex Panican, on stage at the European AI for Finance summit to discuss these questions:

Bibi Ndiaye, Innovation Director and Data Intelligence, BPCE – on AI’s bias in recruitment:

“In the case of Amazon, I’m not sure that the AI is not the problem. The AI used a training set to make those decisions. So if in that training set the human working in HR used to recruit mostly men, you cannot ask the AI to do other things – because the training set uses these examples

If the human makes a discrimination, the AI will do the same.

The risks are very limited in banks, because all AI models being developed are towards the goal of helping humans in making decisions. So in that case we limit the risk.”

Jean-Marc Bonnefous, Co-Founder & Board member, Bonseyes – on AI’s use in sentencing:

“This illustrates all of the risks of deploying AI in a not human-centric way:

It lacks fairness – there’s a bias – and it lacks transparency and explainability; how did you arrive at the decision?

This is typical of what we should watch for when developing an AI application, to ensure we have the right amount of governance in a data driven system so we do not let the system dictate the ethics.”

“Artificial Intelligence is neither artificial nor is it an intelligence. It is a process, and it follows the rules of a process: garbage in, garbage out.”

Frère Eric Salobir, President, Optic Technology – on AI replacing God:

“We see that each technology which is sophisticated enough can appear as being quite magical. We see that there is also a temptation for human beings is to build something and start to worship it – and to expect a kind of salvation from it.

What we have to do is to empower and train people to ensure that they will always consider a machine as a tool. We also pay a lot of attention to situations where human beings are under the control of a machine. More and more we have two kinds of people: People who have machines working for them, and ones who work for machines.

If you work in the warhouse of some of these companies, and all day long you just do what these machines tell you to do, is it good for your dignity? Is it good for your self-esteem? I am not so sure.”

Bonus question: Frère Eric Salobir, on the potential for humans reaching a ‘singularity’ with computers, where we upload our consciousness:

“For the human being could be good, for the species it could be the worst.

There are two ways to extend your life. One is reproduction, which is what we – and the animals and plants – do. The second one is to extend forever your own life, but you remain who you are. So if we decide now that we continue to live instead of having children, instead of having another generation, it also means it would be the end of evolution.

So what some people say “Oh this is so magic”, this kind of post-humanity, but this post humanity would just be standing all the time at the same place, and probably it would not be evolving anymore. And that is not the best for the species.”

Will AI replace God? To see more of what was discussed during the panel, check out the video of our panel at the European AI for Finance conference by startupinside on 3rd of september with the best insights of Bibi NDIAYE, Innovation Director and Data Intelligence, BPCE /Jean-Marc BONNEFOUS, Co-Founder and Board member, Bonseyes / Frère Eric SALOBIR, President, Optic Technology /Alex PANICAN, Head of Partnerships and Ecosystem, The LHoFT