There are deep age, gender, occupation, and class-based social divides in how people use generative AI applications - such as ChatGPT - and how they perceive their impact on job security, new research shows.
People who don't use GenAI and those who are older are more likely to see the technology as a threat to their jobs rather than an opportunity, according to a new survey.
Those who use GenAI frequently in their private lives and at work are more likely to see it as an opportunity than a threat to their job security.
Among those in professional and managerial roles, 41 percent viewed GenAI as a job opportunity. That figure fell to just 25 percent among people working in administrative, sales, or service roles.
In contrast, 39 percent of respondents in admin, sales, and service jobs - and 35 percent of those in manual roles - saw GenAI as a threat to their job security. Among professionals and managers, only 24 percent shared that concern.
Relatedly, concerns are widespread in the working class. 58 per cent of those identifying as working class see AI as a threat, but only 33 per cent among those in the middle class.
Women and older adults were also more likely to perceive GenAI as a threat. Holding a university degree did not appear to be linked to whether someone sees GenAI as a threat or an opportunity.
The study found some associations with political affiliation: Conservative voters more often viewed GenAI as an opportunity, while Reform UK voters more frequently see it as a threat.
The data for the study was collected via a YouGov online survey of 1,061 people during March 2025, and funded by the UKRI as well as Horizon Europe.
"The single strongest factor linked to people's views is whether they actually use GenAI - either at work or in their private lives. Even infrequent users are far more likely to see GenAI as an opportunity. Among non-users, GenAI is viewed as a threat more often.
For instance, 50 percent of frequent users see GenAI as a job opportunity for jobs, but only 17 per cent of non-users. In contrast, only 18 per cent of frequent users see GenAI as a threat to their job security, but 49 per cent of non-users.
Usage varies considerably by age. Only about a third (37 per cent) of 18-29-year-olds said they were non-users, compared to 64 per cent of 30- to 49-year-olds, 78 per cent of 50- to 64-year-olds and 89 per cent of those aged over 65.
Men who took part in the survey used GenAI slightly more than women. 65 per cent of men reported not having used AI, while it is 73 per cent among women. University graduates also reported higher use, with 57 percent of those holding a university degree using GenAI in their private life compared to 75 per cent of those without a university degree.
The survey also asked specifically about the use of GenAI at work. About half of those under 30 report using GenAI at the workplace, though actual usage might be higher. Usage at work declines sharply with age and is also higher among those with university degrees. Usage is particularly widespread among those in professional and managerial roles, followed by those in admin, sales, and service roles, and it is lowest among those in manual roles.
Professor Florian Stoeckel, who led the research, said: "These divides are concerning because they risk reinforcing existing inequalities."
"Our survey also shows major divides in AI usage itself - along lines that may not be surprising, but are still concerning."
