Public concern over artificial intelligence is growing as initial excitement surrounding the technology gives way to increasing fears about data privacy, job displacement and the risks associated with so-called “rogue” AI chatbots. Emerging evidence suggests that even many of the technology’s early supporters are now reducing their use of AI tools or abandoning them altogether.
The Digital Futures Institute at King’s College London, in partnership with the UK AI organisation, has published the findings of a recent public opinion survey showing that 42 per cent of respondents have begun limiting their use of AI chatbots. More than half also believe that the technology poses significant risks.
Researchers from the two organisations warned in the report summarising the survey findings that “people are resisting AI”, highlighting a noticeable shift in public attitudes towards the technology.
According to King’s College London, the most common reason for reducing AI use is concern over data privacy, security and regulatory compliance. The second most frequently cited reason is a preference for continuing to work using existing methods rather than relying on AI tools.
The survey, which included more than 2,000 people across the UK, found that the proportion of respondents who believe the risks of artificial intelligence outweigh its benefits has risen by four percentage points to 52 per cent compared with the previous survey. Meanwhile, the share of respondents expressing an optimistic view of AI has fallen to 34 per cent.
At the same time, many participants expressed growing concern that avoiding artificial intelligence is becoming increasingly difficult. Respondents said they felt uncomfortable with what they perceived as being compelled to use AI-powered chatbots, despite having reservations about the technology.
The researchers noted that most people believe it would be difficult—or even impossible—to avoid exposure to artificial intelligence, even if they wished to do so. They argued that this raises important questions about how users can meaningfully give informed consent to AI use in modern society.
Despite mounting pressure on employees and job seekers to demonstrate both an interest in AI and proficiency in using it, the research team stressed that those choosing not to use AI are not necessarily lacking the skills required to operate chatbot applications. Instead, their decision often reflects personal concerns rather than technical limitations.
While global stock markets continue to reach record highs, fuelled in part by strong investor demand for AI-related companies, executives at major technology firms largely dismiss the prospect of waning enthusiasm for artificial intelligence.
Jack Stilgoe of King’s College London said technology companies often assume that the more familiar people become with new technologies, the more they will embrace them. However, he said the survey findings challenge that assumption, showing that even the strongest supporters of AI have become more concerned over time and are now among its most cautious users.
Kate Devlin, also of King’s College London, said people need meaningful safeguards and effective mechanisms to control their interaction with artificial intelligence. She stressed that users should be able to decide when they use AI, how it interacts with them, and how they can opt out whenever necessary.
