Singapore sets AI risk guidelines as insurers brace for agent claims

Singapore sets the guidelines

MAS, Singapore's financial regulator, has issued its finalised Guidelines on AI Risk Management. MAS expects financial institutions to maintain inventories of their AI use and apply controls in areas including data governance, cybersecurity, testing and human oversight. The guidelines phase in, with sections 3 and 4 taking effect on 7 October 2027, and MAS intends to consult the sector in 2027 on what additional guidance for agentic AI would be useful.

Insurers brace for agent claims

Insurers are bracing for claims worth millions from AI agents that have spun out of control, and Tim Rayner, head of underwriting and claims at Verisk, says the buck stops with the CEO. The Financial Times reports that the personal liability of executives such as OpenAI's Sam Altman and Anthropic's Dario Amodei is now in play too.

Dimon sees a 10-fold jump

JPMorgan CEO Jamie Dimon said threats related to AI "went up 10-fold after Mythos," Anthropic's model. Anthropic, for its part, now reserves the top tier of its Cyber Verification Program, Specialized Access, for verified organizations authorized to test systems such as interbank transfer infrastructure, and partners in its earlier Project Glasswing program uncovered at least 129,000 verified software vulnerabilities between April and July 2026.

HSBC plans deep cuts

HSBC is planning job cuts across its UK wealth management business as part of a broader push to integrate artificial intelligence, the Financial Times reported, and reductions among financial advisers could reach around 70 per cent. The bank plans to cut about half of management and specialist roles in the business, the report said.

Start with the inventory

MAS expects financial institutions to maintain inventories of their AI use. A useful one lists each model and agent, who owns it and what it can reach, and gives a regulator, an insurer and a security team one place to start.