TL;DR

The heads of the CMA, FCA, ICO and Ofcom have jointly written to ministers about personal AI agents, saying several such products are due to launch in the UK and that developers have not all engaged with regulators early or openly. The letter, dated 16 November and published by the DRCF, the regulators’ joint Digital Regulation Cooperation Forum, says firms must meet their obligations before launch and that the regulators “will not hesitate” to use their powers.

The letter went to the Chancellor, the culture and business secretaries and the Minister for AI. It is signed by Paul Arnold of the ICO, Melanie Dawes of Ofcom, Sarah Cardell of the CMA and Nikhil Rathi of the FCA.

What the regulators say

The four chief executives describe personal agents as “a step-change in the way AI is being used”, and say there are real questions about how well the public understands how these systems work. They see benefits too, singling out agents that help consumers save money on everyday purchases.

Their concern is the run-up to launch. “Some firms have engaged early and constructively,” the letter says, “but engagement has not been consistently early or open across the relevant regulators.” Talking to regulators first helps a developer establish whether its service needs authorisation or registration, they argue, and later versions of a product will need testing against the same requirements.

The letter names three risks firms should address: fraud and scams, personal data being shared in ways users would not expect, and misleading recommendations, decisions or information. Agents must also leave users in control and must not undermine the operational resilience of the systems they connect to.

Banks and telecoms firms are in scope too

The regulators also address companies that strike commercial deals with agent providers, naming financial services and telecoms firms. Their own obligations to customers continue to apply when those customers act through an AI agent, the letter says.

Looking forward

The letter follows the DRCF’s joint statement on agentic AI last month, which we covered in our report on the regulators’ personal AI agents statement. Writing to four ministers about the same products a few weeks later suggests, in our view, that the regulators expect launches soon and want government aware of where engagement has fallen short. For firms building or partnering on consumer agents, the practical message is to brief every relevant regulator before going live, not only the one they know best.