Ethics, trust and the rise of AI in Financial Services

For some time, I’ve followed Clara Durodié’s Decoding AI for some time. Decoding AI is a digital publication dedicated to AI in financial services and, in her most recent update, she delves deep into Trust, Ethics and the Rise of AI in Financial Services, which is my specialist subject of the moment.

I won’t repost the whole article – you need to subscribe to Clara’s updates to do that – but I will pick out some highlights.

First and foremost is that AI is now a Board and Executive leadership topic. That kind of seems obvious after my update the other day showing the explosive growth of AI, AI use cases and AI in practice in financial services.

This means that, as with other industries and markets, AI is becoming foundational to financial services. The thing is that AI is not something that is as easy to manage as traditional code, because the code develops and creates its own sense of being. As mentioned yesterday, if you allow AI to run itself then it creates its own societies and groupings. That is not something you can predict, as it is a growth system rather than a rules-based or programmed system.

In financial markets, this is a key issue as we will be applying AI systems to deal with people’s core needs – mortgages, insurance, credit and more – and if the algorithms get these processes wrong, it is not the AI that customers will blame. It will be the financial firm.

Bear in mind that the financial markets depend upon trust and precision, fairness, accountability, and transparency through compliance, governance, and robust internal controls, this is a critical change to the way firms operate. It’s a bit like if you gave pilotless control to an aircraft – you still have to ensure it takes off and lands correctly. The same applies to financial firms. The system has to work but, with AI, work more efficiently and faster.

Clara highlights the risks we face as AI introduces a new class of risks that are latent, embedded in data, and often difficult to detect. These are not just technical challenges. She urges executives to realise that these are ethical questions, and they demand oversight at the highest level.

So, what to do about it? Well, here’s Clara’s recommendation:

“Boards and senior executives must act decisively:

• Demand transparency: Treat AI risk with the same seriousness as financial, operational, or reputational risk. Ensure regular board reporting on ethical impact assessments, model risk management, and incident escalation.
• Integrate ethics into product design: Embed cross-functional collaboration between data scientists, legal, compliance, and ethics experts from the earliest stages of development.
• Align incentives with outcomes: Tie performance and innovation metrics not just to speed and efficiency, but to ethical and customer-centric results.
• Build board-level AI fluency: Improve understanding of how AI systems function, where bias can creep in, and what governance structures are effective. Oversight cannot be meaningful without baseline knowledge.
• Engage with the ecosystem: Collaborate with peers, regulators, and civil society to shape emerging norms and standards. Responsible AI is not just a compliance issue—it is a strategic and reputational one.”

Funnily enough, on that last point, it is a discussion I am taking to EBA Day next week. The theme? How financial institutions, fintechs, and technology providers can co-create ecosystems that are agile, inclusive, and future-ready.

See you there at the Carrousel du Louvre, 27–28 May and, if you haven't already, register here: https://lnkd.in/exsVRis8 to join us.