The International Regulatory Strategy Group ("IRSG") in October published a research paper exploring the opportunity afforded by artificial intelligence ("AI") in the context of the UK financial services industry. The research paper considers how the UK government and regulatory bodies can encourage AI-powered innovation while mitigating any potential risks arising from industry-wide adoption.

About the IRSG

The IRSG is a leading cross-sectoral group for the financial services and related professional services industries which focuses on regulatory developments. Its membership is drawn from senior UK based industry leaders.

Initial conclusions

The IRSG research paper considers AI to be a potential driver of exponential growth and opportunity, noting that financial and related professional services are considered to be among the top three industries that will benefit most from the application of AI.

Concluding that the fundamentals of the UK's regulatory architecture remain robust and fit for purpose for current applications of AI, the report cautions that as technology becomes more transformative and breaks new ground further thinking on policy might be needed. Their recommendation is that industry best practice should be used to leverage existing regulatory frameworks as the most effective way to ensure that there are proper consumer safeguards and market stability, while at the same time facilitating innovation and growth.

Comments on best practice for AI-powered innovation

When commenting on best practice, the research paper draws out the following key areas of focus for businesses.

Fairness, transparency and consumer protection

Firms should develop ethical AI frameworks and focus on the development and use of auditable AI systems underpinned by robust checks and controls, including periodic sampling of end outcomes to build confidence and reduce the risk of unintended consequences.

Data ethics

Firms were encouraged to establish a clear set of data ethics principles for the collection, processing, aggregation and sharing of customer information to build digital trust. Risk frameworks should be developed to incorporate contingency plans for incorrect outcomes associated with a range of inputs, including the use of inaccurate historic data.


Firms should aim to improve digital capabilities/tech literacy across the organisation and at board level to ensure accountability. A shift in corporate oversight away from the traditional approach of delegating tech matters solely to a Chief Technology Officer figure was needed. Roles and responsibilities for AI needed to make sense within existing accountability frameworks to ensure that the application of AI did not challenge traditional accountability lines.

Ecosystem resilience

Firms should ensure there is strong traceability of data and algorithms and develop robust business continuity and resilience plans in the event of a failure/threat. This will require the retention of greater human oversight of AI systems to reduce risk.

The research paper notes that:

Digital transformation with AI means thinking big, starting small and acting faster than before....... Businesses must balance speed, agility and scale. This requires them to focus on the 'brilliant basics' to get AI right: they need to become strategy- and value-driven, put in place strong governance controls, and align their talent management strategy to the broader business objectives.

Much work still to do

the IRSG research paper chimes with many of the concepts emerging from recent UK, EU and other global initiatives (click here for our recent round up report). This is a clear area of developing regulatory focus but much of the output remains at the level of abstract principles. Equally, at an industry level, despite significant momentum in AI adoption, the IRSG research noted that that the maturity of AI solutions varied widely depending on the sector and the type of customers served by firms. Many AI projects remained at the prototyping/proof of concept stage - firms are still experimenting. Only a few financial institutions were taking their first steps towards scaling AI across their organisations. Accordingly, there remains much work to be done before business opportunities are fully realised and the robustness of developing regulatory approaches are truly tested