Picture of Calypso Harland

Calypso Harland

February 6, 2024

Rise & Shine: Expert Tips for Fintechs Navigating AI Regulation

In a packed room at last week’s Rise & Shine event, the spotlight was on the future of AI regulation, unravelling strategies for Fintechs to futureproof their approach by learning how to cultivate trust and manage risk in what is relatively unchartered territory.

Douglas Mackenzie, an expert moderator from FF News | Fintech Finance, skillfully guided panels that delved into the ethical dimensions of AI in UK fintech and strategies for effective risk mitigation.

The event featured an exploration of the EU AI Act delivered by Michael Charles Borrelli, . Alexandra Matthews, representing Munich Re, shared an invaluable perspective from the insurance realm, unveiling practical tips for fintechs to guard against potential risk.

Kate Shcheglova-Goldfinch, Andrei Lebed of Koodoo, and Ken Cassar of UMNAI joined as panellists whose contributions added depth to the discussions.

Here are their top tips on the evolving landscape of AI regulation and its profound implications for the fintech industry.


Top Tips from our AI Experts

Douglas Mackenzie
Moderator: Douglas Mackenzie
Douglas Mackenzie is the Chief Content Officer at FF News / Future Finance. A highly-regarded journalist, executive producer, and expert moderator in Finance.

“Keep Calm and Carry On: Despite the incredible headlines, ultimately AI is a tool to be utilised for specific tasks. The industry seems to be caught up in the hype, applying potential use cases to every element and as a result, the risks seem far greater than what they might be.”

“Understand what risks you would be uncomfortable with and mitigate them: With the introduction of this incredible tool, understand what elements of the risk would be the most damaging to your organisation and either work to reduce it, through hiring specific AI C-Suites or even transfer that risk to an insurance organisation.”

“Start Now: Educate now, understand your competition will be comfortable with using AI and ask yourself, how will your organisation fare against them in half a decade?”

Michael Borelli Medium (1)
Speaker: Michael Borrelli
Michael Borrelli is a Director at AI & Partners, specializing in compliance and regulation. With a focus on fintechs, he works to audit and assess AI systems, ensuring EU compliance.

“Establish a Risk Management System: The EU AI Act mandates the establishment of a risk management system for high-risk AI systems. This system is a continuous, iterative process that runs throughout the AI system’s lifecycle.”

“Testing and Compliance: High-risk AI systems are required to undergo testing to identify the most appropriate risk management measures.”

“Consideration of Vulnerable Groups: The Act emphasizes the need to consider the potential impact of high-risk AI systems on vulnerable groups, including individuals under the age of 18.”

Kate Schl
Speaker: Kate Shcheglova-Goldfinch
Kate Shcheglova-Goldfinch is the Programme lead for Regtech and AI governance at the Cambridge Centre for Finance, Technology & Regulation. Her research focuses on Ethical AI principles in finance.

Strategic Compliance Focus: “The transition of the AI sector to maturity demands a strategic approach to compliance. Elevating AI issues to board-level priority is paramount, requiring a clear understanding of AI model specifics and potential regulations.”

Organisational Preparedness:“To navigate the evolving AI landscape, businesses must prioritise a proactive stance. This involves monitoring regulatory updates, considering external advisers and AI-driven Regtech solutions, establishing a CAIO position, crafting an AI policy, and aligning ethical principles with regulations.”

Innovation and Risk Mitigation: “Collaboration with regulatory bodies and testing AI models in innovative legal frameworks like digital sandboxes are crucial steps. Businesses should explore these avenues to mitigate risks and ensure compliance through innovative regulatory frameworks.”

Andrei Koodoo headshot (1)
Speaker: Andrei Lebed
Andrei Lebed is the CEO and Co-Founder of Koodoo, a company recognized for its innovative applications of AI and technology in the UK mortgage sector.

“When I evaluate the risk of AI applications, I look at it on a two-dimensional spectrum, taking into account the risk of the AI giving the wrong answer, and the level of human intervention in the design.”

“Be careful of using customer chatbots in front of customers without serious testing – recent examples of chatbot hackers are a concern for any financial services institution.”

“Keep an eye on your model diversity in 2024 – leading LLMs have been trained on mostly English-language data sets, which in itself contain cultural biases that we might be blind to.”

Speaker: Ken Cassar
Ken Cassar is the co-founder and CEO of Umnai, a deep tech AI company that recently launched Hybrid Intelligence, a ground-breaking, Neuro-Symbolic AI framework.

“Be mindful that leveraging data through AI in multiple processes in your business is likely to be table stakes for remaining relevant and competitive in a 3 to 5 year time window.”

“Prefer AI frameworks that give you oversight control over the AI models you deploy and the outputs they produce. Explainability, traceability, and agency over the AI outcomes are critical.”

“Be sensible in your approach to AI. Don’t try to boil the ocean. Pick a use case where your data is qualitatively and quantitatively good and you know the use case well. Start small and increment your project.”

Alexandra Matthews
Speaker: Alexandra Matthews, Munich Re
Alexandra Matthews is an Underwriter of AI Performance and Liability Risks at Munich Re, one of the world’s largest reinsurance companies. Munich Re is the first insurer to absorb legal and contractual liabilities arising from the implementation of AI.

“No matter how well an AI is trained, it will always make errors. This is a risk that you cannot remove, but you can minimize it by transferring the risk via insurance to offload some potential liabilities off your own balance sheet.”

“Articulate the risk: is it a risk that underperformance will cause you or your users a financial loss? Or is the risk that your AI model makes a correct prediction but infringe upon someone’s IP rights or discriminates? Knowing what AI risk is likely to have the biggest impact on your business is a valuable start to de-risking user operation of AI.”

For those eager to continue the dialogue, a new AI discussion channel has been launched online at Fintech Fringe Forum.

The next Rise & Shine event will focus on supporting Female Founders in Fintech with Fundraising on March 07. Please contact Calypso Harland, if you would like to be involved.