Most Britons won’t give AI much control, survey finds

A new survey of 2,000 UK adults finds strong public limits on AI agents that can book trips, manage accounts and make purchases. The biggest worries are hacking, mistakes and unauthorized actions, and many respondents want humans to keep final control.

Why it matters: - AI agents are moving toward tasks that affect money, accounts and daily decisions. - The survey suggests most UK consumers still want human oversight, especially when AI can act independently. - That matters for companies rolling out AI in banking, insurance, utilities and other trust-sensitive services.

What happened: - Eskenzi PR commissioned a nationally representative survey of 2,000 UK adults. - Censuswide conducted the research between 16 and 18 September 2026. - 84% of respondents said they would only give an AI agent moderate or low control over day-to-day life, or no control at all. - 34% said they would not be comfortable giving an AI agent any control. - Only 11% said they would be comfortable giving AI high or very high levels of control.

The details: - 27% would let an AI agent prepare things for them but would want to approve every action. - 23% would allow AI to handle routine tasks independently but would still require permission for other actions. - 36% worried their data could be hacked or stolen if an AI agent acted on their behalf. - 31% worried an AI agent could make a mistake. - 30% were concerned an AI agent could access too much personal information. - 28% worried an AI agent could spend money or make an unintended purchase. - 22% were concerned an AI agent could communicate with someone without their knowledge. - 14% worried an AI agent could implicate them in something illegal. - That concern rose to 21% among 18- to 24-year-olds. - Jonathan Armstrong, partner at Punter Southall, said people often assume AI is a legal “Wild West,” but users are usually responsible for the AI they use and the consequences. - 47% of over-55s said they would not be comfortable giving an AI agent any control over their day-to-day lives. - 21% of 18- to 24-year-olds and 18% of 25- to 34-year-olds said the same. - 20% of 25- to 34-year-olds would be comfortable giving AI high or very high levels of control, compared with 5% of over-55s. - If a bank, insurer or utility used an AI agent to advise its board on major decisions, 37% would want reassurance that a human always had the final say. - 31% would want to know exactly which decisions the AI was involved in. - 24% would want to know how personal data was being protected. - 20% would consider switching provider. - 18% said they would be less likely to buy additional products or services from that company. - 10% would see the use of AI as a positive sign the company was modern and embracing technology. - 6% said it would make them more likely to use the company. - If an AI agent with access to user accounts was hacked, tricked or carried out an unauthorized action, 40% said the AI maker should be primarily responsible. - 25% said the company or service the AI was accessing should be primarily responsible. - 18% said the individual using the AI should be primarily responsible. - Yvonne Eskenzi, co-founder of Eskenzi PR, said people appear willing to use AI for routine tasks but want visibility, accountability and human control over consequential decisions. - On broader AI risk, 22% said they were more worried about what humans would do with AI than AI itself. - 19% said AI should be treated as a genuine threat to humanity. - 18% said the danger is real, but extinction by 2030 sounds far-fetched. - 15% said AI could cause serious problems but would not wipe out humanity. - 9% said they were not worried about AI at all. - Censuswide said it is a member of the Market Research Society and British Polling Council and follows MRS and ESOMAR standards.

Between the lines: - The findings point to a clear comfort gap between AI as a helper and AI as an autonomous decision-maker. - Consumers appear more willing to use AI for preparation and routine work than for purchases, communications or account access. - The accountability results suggest public trust will depend heavily on who bears responsibility when AI systems go wrong. - The age split shows younger adults are more open to delegation, but even they want limits.

What's next: - Companies using AI agents will likely face stronger demands for human oversight, clearer disclosures and tighter data protections. - Organizations that give AI more authority may need to prove where the AI is used and how liability is handled. - The survey’s findings can be read in the full report.

The bottom line: - Britons are not rejecting AI outright, but they are not ready to hand it the keys, either.