The Pew Research Center, a nonpartisan think tank that tracks public opinion, released a report today on how workers feel about AI.
The technology has become an increasingly common workplace fixture over the past few years. And its role is likely to grow as AI becomes more capable, thanks to advances such as the large language models, like GPT-4, that gave us ChatGPT and a growing number of other tools.
While there’s no shortage of reports about people’s attitudes toward AI, Pew’s data is sizable and relatively fresh, drawing from 11,004 US adults who were consulted between December 12 and 18 of last year—just as ChatGPT mania was taking hold after its release at the end of November.
The report suggests that most workers expect AI to transform hiring, firing, and evaluations. Many people report feeling uncertain about what those changes might look like, and concerned about AI’s potential impacts.
Some 68 percent of those polled said they expect AI to have a major impact on jobholders over the next 20 years. Curiously though, only 28 percent said they thought AI would affect them personally, while 38 percent were unsure what the outcome might be for their own work.
Those responses reflect the fact that nobody really knows how AI will change jobs and work in the coming years. The technology is evolving quickly, and its impact often differs greatly between industries and even roles.
We can, however, expect existing uses of the technology to expand and become more sophisticated. Some employers already use AI to help screen job applicants, while enterprising job hunters seek to outwit the algorithms with clever tricks. In theory, AI technology has the potential to make hiring fairer and increase workplace diversity. But in practice it has sometimes done the opposite, leading the US government to warn employers about the potential for algorithms to discriminate against people with disabilities.
The Pew survey reflects this contradictory picture, with 47 percent of people saying they think AI would do a better job than a human in hiring but 41 percent opposing the use of AI in hiring.
Workplace surveillance is an area of general concern with 81 percent of those polled saying that more use of AI will lead to workers feeling inappropriately watched.
As WIRED has previously reported, many workers now have to work under the supervision of AI-powered surveillance systems that monitor productivity and flag misbehavior. Some truck drivers and warehouse staff are monitored by cameras and algorithms, and a rising number of office workers have their keystrokes and mouse clicks logged and analyzed. This kind of data collection is generally intended to improve productivity, but it can be dehumanizing and demoralizing and push employees into working too hard or taking unnecessary risks, like driving trucks too fast.
Fear and doubt about AI’s role in workplaces is understandable because we are at something of a crossroads. Some economists argue that the designers of AI systems need to think about how their products will affect workers and aim to enhance their output rather than replace them. Similarly, employers can do more to ensure that the tools they use to track workers do not cause undue harm.
So far, the companies developing these AI tools and the employers that use them have not risen to the moment, often placing workers at the mercy of systems whose benefits are in question. It’s time to stop and think about how workplace AI is used, a need that will only grow more pressing as the technology advances.
