缅北禁地

28 May 2026

A new report from the London Foundation for Banking and Finance (LFBF) and the 缅北禁地 (IFoA) introduces a new framework for understanding AI risks in financial services. It acknowledges that generative AI presents opportunities as well as risks but points to 鈥榰ncomfortable tensions鈥� 鈥� where the same features that make the technology valuable also make it difficult to govern, explain, trust or contain.

鈥業t鈥檚 still not magic: Framing the risks facing financial services in the Gen AI era鈥� builds on its . As part of the research, we surveyed senior financial services practitioners and observers and found that: 

• 70% agreed 鈥榬isks arising from the use of AI are among the greatest risks facing my sector over the next five years鈥�.
• 75% agreed 鈥榯he risks posed by AI to my sector have increased substantially since generative AI technologies have become widely available.
• The top three risks highlighted were cyber threats, misleading outputs, and knowledge gaps.

Generative AI has fundamentally changed the risk landscape not only because it can 鈥榟allucinate鈥�, or present false or misleading information that appears authoritative. It also makes AI widely accessible, persuasive, easy to use and embedded in every day financial workflows. As firms increasingly build AI into tools and infrastructure, many of the most complex risks are ecosystem risks. Decisions that appear sensible for individual firms can create hidden dependencies and shared points of failure across the financial system. 

The report鈥檚 AI risk framework identifies nine risks which are grouped into three broad categories: 鈥榦utcomes鈥�, 鈥榦perating environment鈥� and 鈥榮ystem鈥�. These provide a useful way to trace how AI risk moves through the financial services ecosystem: from the outcomes experienced by customers and society, to the environment in which firms deploy AI, to the system-level dynamics through which risks can scale and spread.

The framework treats AI risks as trade-offs, rather than standalone downsides. Often, the risks are the direct counterpart of AI鈥檚 benefits 鈥� sharper prediction, deeper personalisation, greater scale and complexity, and more autonomous decision-making. Although risks cannot be eliminated altogether, the framework outlines how many can be moderated through better risk management, governance, and regulation.

Keyur Patel, LFBF Research Associate and report author, said: 鈥淭he same characteristics that make AI useful in financial services also create many of the risks that make it so difficult to govern. The hard question, then, is not just whether these risks can be mitigated, but how much risk we are willing to live with in exchange for the benefits. That is why a recurring theme in this report is 鈥榰ncomfortable tensions鈥�: the same machinery can widen inclusion and sharpen exclusion; 鈥榟uman in the loop鈥� is not the same as human control; and concentration is baked into how AI systems are built. Generative AI gives these tensions new force. It lowers barriers to use, makes AI feel relatable and trustworthy, and is increasingly embedded in how financial institutions think and work. That matters because AI outputs can be useful, confident and wrong at the same time 鈥� and 鈥榤ostly right鈥� can be dangerous.鈥�

Paul Sweeting FIA C.Act, IFoA President, said: 鈥淎I is a defining force of our time. The IFoA鈥檚 Artificial Intelligence and Emerging Technologies Practice Board is exploring how transformative technologies are reshaping actuarial practice and influencing broader societal systems. It is also exploring what we need to do to seize the opportunities and manage the risks associated with AI-adoption. With our unique combination of technical skill, communication and professional oversight, actuaries must play a key role making sure that AI is working as it should.鈥�

~ENDS~