84% in the UK Would Not Give Autonomous AI Systems Broad Authority

    A new survey conducted in the UK has found that people are not ready to give AI agents extensive control. “Censuswide” surveyed 2,000 adult UK residents. An AI agent is a system that independently performs tasks on a user's behalf. Of the participants, 84% want to give such systems only moderate or low-level access, or no control at all.    Of those surveyed, 34% said they would not give an AI agent control over any area of their lives. Another 27% want the system to prepare the task but wait for approval before carrying it out. A further 23% are willing to entrust routine tasks to an AI agent but prefer to retain the final say on important tasks.    Only 9% of participants have no concerns about AI. Of the respondents, 19% consider artificial intelligence a threat to humanity. The risk of an AI agent committing fraud or another illegal act also drew attention in the survey. According to 40% of participants, if the system gets out of control, the company that created it should be held responsible.    Among those over the age of 55, 47% do not want to give an AI agent control over their lives. In the 25-34 age group, however, 20% are not concerned about giving the system a high level of control. Overall, 14% of participants fear that using AI could expose them to legal problems. Among those aged 18-24, this figure is 21%.    According to lawyer Jonathan Armstrong, AI is not an area that falls outside the law. In most cases, the person using the system will be responsible for any illegal acts it commits and their consequences. Even if an AI agent takes an unexpected action, the user may be held responsible. It is therefore important to understand the potential risks before enabling automated tasks.