Bank of England calls for agent-specific AI rules

A Bank of England official said existing AI rules don't fit systems that act on their own in financial markets, per Finextra.

For fintech builders already treating AI compliance as covered, that box won't stay checked. The target is shifting from AI features sitting on top of a product to the agents making decisions inside it.

The reasoning tracks with where regulators say the actual risk lives: in the agents themselves, not the AI systems generally. That's a narrower target than blanket AI guidance, one that puts autonomous decision-making front and center instead of treating agents as just another AI feature.

What the report doesn't say: how soon a bespoke framework might land, or what specific obligations it would put on agents beyond today's AI rules. The official flagged the mismatch, not a timeline.

My bet: the first bespoke agent rules target systems that act on their own inside the market, not chatbots wearing an AI label. I'd watch for other central banks to follow with their own agent-specific frameworks, instead of stretching current AI guidance to cover it.

This lands while agentic systems are already live inside markets, ahead of any framework written specifically to govern them, per Finextra.