2023 is the year that AI stopped its walk to sprint into the spotlight, especially in the finance world. I had the pleasure of diving into this tech tsunami at the AI for Finance 2023 Summit in Paris, alongside our Deputy CEO Alex Panican. There, I soaked up the wisdom of industry trailblazers, eager to share the synergy between generative AI and finance – a combination pushing the boundaries of innovation and efficiency.
The blueprint of AI-led strategy
Deutsche Bank’s AI journey
Gil Perez, Chief Innovation Officer at Deutsche Bank, emphasised that successful AI implementations require a keen focus on regulatory compliance, with data centres audited and approved by regulators. The company’s three-year strategic plan around generative AI, in partnership with Google and NVIDIA, highlights the growing “AI/ML” focus.
Google Cloud’s Perspective
Laurence Lafont, Vice President of Strategic Industries EMEA, pointed out the significant demand for generative AI across organisations. The emphasis is on safety, privacy, and capabilities like creation, summarisation, automation, and discovery. Google aims to be an “AI-first” company, investing in “Bold and Responsible” AI frameworks such as the Vertex AI platform.
Generative AI’s Multifaceted Role
Generative AI is versatile, enhancing fields like risk management and offering up new unique value propositions. It’s predicted to soon include hyper-realistic avatars and integrate emerging technologies like quantum computing.
Aligning AI with Business Goals
Philippe Vallée, Executive Vice-President of Digital Identity and Security at Thales, and Mung Ki Woo, Director of Sopra Steria, discussed AI’s role in improving user experiences in banking. Thales focuses on “True AI” – Transparent, Understandable, and Ethical AI, stressing explainability and GDPR compliance. The company collaborates with cloud providers for data sovereignty and security, also aiming at financial inclusion.
AXA’s AI Integration
Alexander Vollert, AXA Group COO & AXA Group Operations CEO, highlighted AI’s core role in their strategy, with a focus on training and data preparation. AXA also collaborates with Microsoft for a secure GPT solution, emphasising AI’s potential to streamline operations and set ethical standards.
Lingua Custodia’s Financial Document Processing
Olivier Debeugny, CEO of Lingua Custodia, participated in the AI for Finance summit as an exhibitor. As a member of our ecosystem at the Luxembourg House of Financial Technology (LHoFT), Olivier shares his experiences and the various endeavours pursued by his company:
“Lingua Custodia is the expert in Language technologies and Generative AI applied to the financial domain. On our Verto secure platform, our clients can perform financial document processing tasks such as machine translation, speech-to-text, data extraction, and retrieval augmented generation.
Our sales, product, and tech teams were very pleased to attend the AI for Finance forum this year, particularly because we had several important developments to announce that we feel truly add value to financial document processing:
Firstly, the rollout of a multilingual document analyser that can help our clients retrieve key information from large documents to answer questions on in-house procedures or product offerings such as Due Diligence Questionnaires or RFPs.
Secondly, the creation by our experienced research team of a Large Language Model fine-tuned by Lingua Custodia on financial sentiment analysis which performs better than bigger models such as GPT4 or BloombergGPT.
This event was therefore the opportunity for us to illustrate to our clients that there is an alternative to big US names when it comes to Generative AI. Following several interesting discussions with the other attendees, we also realised that financial institutions were not all at the same stage of their journey with this technology. While some of them already have GenAI tools in production, others seem to think that this is just the new hype like the blockchain or the Metaverse.
This event gave us the opportunity to explain and demonstrate that GenAI is already present and used by millions of people. The rapid adoption of this technology within the banking sector is essential for its community to remain attractive to both clients and employees.”
Transformative Potential: Bridging Technology and Regulation
AI’s Importance in Finance
Emma Sezen, Director of AI for Finance, Artefact Open Innovation, and Manuela Veloso, Head of AI Research at JPMorgan Chase, highlighted generative AI’s game-changing role in finance. Generative AI models function based on the likelihood of sequences, providing a foundation for language understanding. However, their effectiveness depends on the quality of training data.
Regulatory Challenges and Trust
The complexity of AI models, particularly in finance and healthcare, necessitates explainability and human oversight. Traditional regulation methods are inadequate; a shift towards outcome-based regulation and testing is needed. Companies should consider the balance of public and private data in training AI models and analyse variances in responses from large language models.
Société Générale’s AI Approach
Lorenzo Croati, Partner of Artefact Open Innovation, and Aurore Gaspar, Deputy Head of Société Générale French Retail Banking Network, discussed Société Générale’s AI integration in their retail banking model, focusing on client satisfaction and a data-driven strategy. The bank utilises AI for client identification and to assist relationship managers in accessing information.
Conclusion: Engineering Trustworthy AI
Marçin Detyniecki, Head of Research and Development & Group Chief Data Scientist at AXA, and Juliette Mattioli, Senior Expert in Artificial Intelligence at Thales, reminded us that AI isn’t just a tech marvel; it’s about clarity, safety, and fairness. In the fast-evolving AI landscape, ethical practices and bias mitigation are non-negotiable. Companies face a crossroads: build AI in-house or seek partnerships, but the goal is clear: we need trustworthy AI that garners public confidence and safeguards interests.
Navigating AI in finance involves complex compliance issues, requiring a focus on processes and use cases over the technology itself. Policies should be adapted to integrate AI-related issues, improving legacy systems with AI upgrades.
Wrapping up, this summit was an eye-opener: responsible, reliable, ethical AI is not just a goal; it’s a journey reshaping finance at its core, demanding a deep dive into both tech and context.