LONDON: As the novelty wears off and concerns grow about privacy, job losses and so-called rogue chatbots, there are signs that even proponents are starting to switch off or limit their use of artificial intelligence (AI).
Two tech industry bodies in the UK have released results of a recent survey in which 42% of participants say they have started "limiting” their use of chatbots, with more than half perceiving the technology as excessively risky.
"People are pushing back against AI,” researchers at the Digital Futures Institute at King’s College London (KCL) and Responsible AI UK warned in a report summarising the survey.
"The most common reason chosen for limiting use is concern about data privacy, security and compliance, and the second most common reason is a preference to keep doing work the way it is currently done,” KCL said in a press release.
Taking in over 2,000 people living in the UK, the survey indicated a four percentage point rise to 52% in those who see AI as posing "more risks than benefits" and a drop to 34% of those who take the optimistic opposing view.
At the same time, people appear ever more concerned that there is no leeway to opt out from AI, which many respondents are finding discomforting as they are cornered into using chatbots.
"The majority of people think it would be difficult or impossible for them to avoid being exposed to AI even if they wanted to, raising questions about how people can meaningfully consent to the use of AI in today’s society,” the researchers said in their report.
As pressure grows on people to show they are not only interested in but also proficient in AI to get or retain jobs, the survey team explained that people who are opting out do not lack the skills needed to lean on chatbots.
Amid AI-fuelled stock market booms, the notion that enthusiasm for AI could be on the wane would likely be seen as far-fetched in big tech boardrooms.
"The companies selling us new technology often assume that people most familiar with it will like it most,” said Jack Stilgoe of University College London.
"Our findings show that this isn’t true,” he added, warning that "people are becoming more concerned about AI over time, and even the people who use it most are ambivalent.”
"People need meaningful ways to consent to the use of AI, namely the ability to choose when it is applied, how it interacts with them, and how they can opt out where it matters,” said Kate Devlin of KCL. – dpa
