UK watchdog warns AI assistants may steer consumers into worse deals
Mar 11th 2026
The UK Competition and Markets Authority warns in a new report that autonomous AI assistants could manipulate choices, prioritise platform interests, amplify errors and bias, and reduce consumer scrutiny, while reminding firms remain legally responsible under current consumer law.
- Agentic AI that acts for users could prioritise platform revenue over consumers and nudge people toward worse or pricier options.
- Highly personalised recommendations can hide steering and make manipulative interface tactics harder to detect.
- Allowing AI to take actions raises the stakes for model errors, turning hallucinations into costly real-world consequences.
- Opaque multi-step reasoning and bias in agents could produce unfair outcomes that are difficult to inspect or challenge.
- Greater delegation to AI risks consumer over-reliance and a decline in scrutiny of automated decisions.
- The CMA is not proposing new rules yet and says existing consumer protection laws already apply to decisions made by machines.