In January 2025, we hosted a 2-day artificial intelligence (AI) Sprint. Find out how we are using feedback from this event and how we want to continue working togethe
In January 2025, we hosted a 2-day artificial intelligence (AI) Sprint. Find out how we are using feedback from this event and how we want to continue working togethe
Jessica Rusu comment on the AI Sprint
The AI Sprint discussed the opportunities and challenges of AI in financial services. 115 participants took part from across industry, academia, regulators, technology providers and consumer representatives.
In teams, they discussed how AI may develop in financial services over the next 5 years, and the FCA’s role enabling firms to embrace the benefits of AI while also managing the risks.
We summarise the feedback received alongside the responses to the AI Input Zone. We are reviewing this feedback to consider how we can further support firms looking to adopt AI in a safe and responsible way and encourage beneficial innovation and growth.
The AI Sprint is one component of our AI Lab[2], which continues to support firms’ implementation of AI use cases and provide a pathway to engage with firms and wider stakeholders about AI-related insights, discussions and case studies.
AI Sprint
Blue line between chapters
During the AI Sprint, 4 common themes came from participants’ discussions and suggestions:
Participants highlighted the importance of firms understanding how existing regulatory frameworks apply to AI. Teams suggested areas where the FCA could clarify, or build on, existing requirements to help firms understand regulatory expectations and to support beneficial innovation.
Participants saw trust in AI as vital for its successful adoption. Teams discussed that if firms and consumers felt able to trust AI, then they might utilise it more, including seniors buying into new AI use cases and consumers engaging with new offerings. Participants agreed that without trust, the full benefits of AI in financial services wouldn’t be realised.
Participants emphasised that all parties involved in AI needed to work together to develop solutions. This includes domestic and international regulators, government, financial services firms, academics, model developers and end users.
Blue line between chapters
Participants appreciated the need for a safe testing environment to encourage responsible innovation. Suggestions included using the FCA’s sandboxes and innovation services to create this safe space, as well as providing access to datasets for innovators to develop and improve AI solutions.
This Sprint focused on 2 main phases for participants to explore:
Teams first considered how AI might develop in financial services and how current use cases may change and emerge. Findings included:
Participants discussed virtual AI assistants and AI financial advisers. They explored how advancements in agentic AI (ie autonomous models) could help personalise and scale-up consumers’ interactions with financial services, leading to better outcomes.
Participants highlighted opportunities to increase the efficiency and effectiveness of internal processes. This included using AI to automate regulatory compliance, enhance fraud detection and help customer support agents.
Use cases: AI financial advisers, internal process automation, compliance tools, emotion AI and personalisation, agents.
Teams then discussed what conditions might be needed over the next 5 years to enable safe and responsible AI adoption. These included:
Participants discussed adopting a measurable way of assessing whether AI should be used in certain circumstances. For example, AI use should ‘at least be held to as high standards as a human process’. This was linked to potential preconceptions of mistrust in AI systems.
Participants discussed how essential clear model, data, cloud and technology foundations are for AI innovation. Upskilling staff on AI and effective internal governance were also highlighted as requirements for success.
Collaborating to develop common data standards and definitions could enhance interoperability, which would create more opportunities for innovation. Transparency and data sharing were highlighted as crucial to achieving this.
Factors that contribute to safe AI adoption: measurable success criteria, model/data/cloud/tech foundations, staff upskilling and internal governance, common standards and interoperability.
Teams then looked at topics related to the current regulatory framework for financial services, which were split into 3 overarching themes.
Blue line between chapters
We are committed to building on the momentum from the AI Sprint to support the safe and responsible adoption of AI across UK financial services.
The feedback and insights received during the AI Sprint – as well as on the AI Spotlight and AI Input Zone – have provided valuable information. They are being used to shape our future work on AI in line with our statutory objectives.
Launching the Supercharged Sandbox, offering innovators greater computing power, infrastructure, datasets, and mentorship to support the testing and validation of AI solutions. The programme is being shaped by insights from the Sprint and identified focus areas to address key adoption challenges. The goal is to equip innovators with the necessary tools, expertise, and regulatory engagement to drive AI adoption in financial services.
Considering whether there are areas of uncertainty where regulation could be restricting safe and responsible AI adoption.
For example, one identified area of uncertainty was around data protection and privacy. With the ICO, we have since written to trade associations[4] to obtain more information about these persistent challenges and how we can better support industry in tackling them.
Ensuring we effectively influence the international standard-setting bodies work (eg IOSCO, FSB, GFIN) on AI to ensure we can support safe and responsible adoption in the UK.
Engaging bilaterally with other regulators to explore cross-cutting considerations as well as collaboration opportunities on specific themes.
Providing clear communication about our approach to AI and continuing engagement with stakeholders through our AI Lab[2]. The AI Spotlight will run on an ongoing basis, where new themes will be added based on the use cases discussed in the Sprint.
Links