This week, the UK government took its most concrete step yet toward AI-powered finance, publishing a landmark AI Adoption Plan for financial services while Aviva’s AI fraud figures revealed the threat is already materialising at scale. Across the Atlantic, a joint report co-authored with Anthropic and OpenAI exposed a systemic ‘silent AI’ liability time bomb inside conventional insurance policies — a structural risk now forcing underwriters to rewrite the rules of coverage.
Top story: HM Treasury publishes its Financial Services AI Adoption Plan, accepting all ten recommendations to accelerate safe AI deployment across UK banks, insurers and asset managers.
HM Treasury Issues UK Finance AI Adoption Roadmap With Ten Mandates
GOV.UK · Regulation
On 14 July 2026, HM Treasury published its Financial Services AI Adoption Plan, developed by independent AI Champions from Starling Bank and Lloyds Banking Group. The plan spans five strategic areas — regulatory framework, AI-powered financial advice, resilience, skills, and agentic payments — and calls for a new trust framework to govern autonomous payment agents, including liability rules and identity checks. Treasury accepted all recommendations and confirmed it will work with the FCA and industry on immediate next steps, marking a significant shift from principles to implementation for UK financial services firms.
https://www.gov.uk/government/publications/ai-adoption-plan-financial-services
90% of Insurers’ AI Exposure Is Hidden In ‘Silent’ Policies, Report Warns
PYMNTS · Risk
A landmark report from the AI Underwriting Company, co-authored with researchers from Anthropic and OpenAI, found that more than 90% of insurers’ exposure to AI risk is embedded in ‘silent’ cover within conventional policies — meaning the risks are largely unpriced and in many cases unnoticed. The research flags a structural shift: AI agents that take autonomous action (rather than simply generating responses) are creating professional negligence and wrongful death exposures that existing policy language was never designed to handle. One in five insurance professionals surveyed by Gallagher in 2026 reported their clients had already suffered losses linked to AI risk.
https://www.pymnts.com/insurance/2026/insurance-industry-may-be-unprepared-for-agentic-ai-risks/
Aviva Reports £233m in AI-Fuelled Suspect Claims Across Its Brands
Life Insurance International · Risk
New data from Aviva reveals the insurer flagged more than 18,000 suspect claims worth £233m across its brands in 2025, with AI identified as a key contributing factor enabling faked car crashes, exaggerated claims, and falsified medical documents. A GlobalData survey of UK commercial insurance brokers found that AI now ranks as the fourth-greatest threat facing the sector in 2026, up sharply from eighth place in 2025, with 9.6% of brokers now naming it their single biggest concern. The data signals that AI-enabled fraud has moved from hypothetical risk to a live and escalating P&L issue for UK insurers.
Freshfields: UK Treasury Puts Agentic Payments Governance at Centre of AI Plan
Freshfields · Regulation
Legal analysis from Freshfields highlights that a distinctive feature of HM Treasury’s new AI Adoption Plan is its focus on agentic payments — autonomous AI systems capable of initiating financial transactions. The plan calls for an upcoming Treasury consultation to establish a dedicated framework covering liability allocation, identity verification for autonomous agents, and authentication standards. The report also accelerated the UK’s Critical Third-Party regime, with four major cloud and AI providers designated as CTPs from 13 July 2026, directly affecting how banks and insurers manage AI vendor concentration risk.
US Treasury Drops AI Risk Framework and Glossary for Financial Sector
U.S. Department of the Treasury · Regulation
The US Treasury released two new resources — a common AI terminology glossary and a tailored risk management framework — designed to standardise how financial firms identify and govern AI risk. The resources are positioned as part of the President’s AI Action Plan and aim to accelerate responsible adoption while reinforcing systemic financial stability. For US and internationally active banks and asset managers, the framework sets expectations that examiners are likely to reference during supervisory reviews, closing a gap that the SEC and Federal Reserve had previously flagged around AI governance documentation.