AI Live Testing: How it can support safe and responsible AI deployment
AI Analysis
The FCA's AI Live Testing initiative provides a voluntary, structured program for firms with mature AI proofs-of-concept (POCs) to test AI-driven services in controlled real-world environments under regulatory oversight and support from technical partner Advai. This matters because it enables safe progression from 'POC paralysis' to deployment, while helping the FCA gather insights on translating AI principles into consumer and market protections, informing future regulation. Participation enhances firms' governance, risk management, and evaluation frameworks for responsible AI use in financial services.
Key dates
- October 2025
- - First cohort began testing (historical reference)
- 19 January 2026
- - Second application window opens
- 2 March 2026 Deadline
- - Application deadline for second cohort
- April 2026
- - Testing starts for second cohort
- Mid
- March 2026; - Notification of successful applicants
Suggested considerations
- Review FCA's Terms of Reference (PDF) for eligibility, focusing on mature POCs and enterprise-level AI systems.
- Submit application form via FCA portal by 2 March 2026 if ready for live testing; contact suptech@ fca.org.uk for queries.
- Prepare documentation on AI system components (model, context/risks, governance, human oversight, evaluation, controls) for three-phase process.
- Assess internal governance, data, risk frameworks, and monitoring for AI readiness; consider non-participation but monitor for future FCA expectations.
- Firms not selected should use insights from first cohort (e.g., evaluation frameworks) to strengthen internal AI practices.
What changed
This is not a mandatory regulatory change but a voluntary testing service launched by the FCA; no new enforceable requirements are imposed. Key elements include a holistic focus on the AI system (model + deployment context, risks, governance, human-in-the-loop, evaluation, input/output controls) rather than isolated foundation models. The program features three phases: Discovery, Framework validation, and AI system testing (quantitative/qualitative), emphasizing live monitoring, governance, and risk management. It complements the FCA's Supercharged Sandbox for earlier-stage AI exploration.
Compliance impact
Urgency: Medium - Voluntary program, but signals FCA's proactive stance on AI oversight; non-participation risks lagging in best practices for Consumer Protection / Conduct and Operational Resilience / Outsourcing as regulator builds evidence for potential rules. Matters for competitive edge in AI deployment and demonstrating alignment with principles-based regulation amid 'POC paralysis'. Early m
Who is affected
Related regulations
References
AI-generated analysis. May contain errors or omissions — verify with the original FCA source before acting. Full disclaimer.
What the FCA said
AI Live Testing now open for applicationsAt the FCA, we’re providing a structured but flexible space where firms can test AI-driven services in real-world conditions, all with our regulatory support and oversight and help from our technical partner, Advai. Collaboration and communication is at the heart of what we are…
Extract from FCA . Read the full notice at the source for the authoritative text.