Do artificial intelligence instruments akin to ChatGPT remove jobs, create new ones, or each? There’s plenty of hypothesis relating to this query, however little strong knowledge within the U.S. to reply it.
I’m a sociologist who researches AI, among other topics. Searching for solutions to that query and others, I commissioned a web-based survey by YouGov, a market analysis company, to ask 1,250 U.S. employees between July 30 and August 4, 2026, about their experiences with AI and their attitudes towards the know-how.
The survey coated employees’ present use of AI, their perceptions of the way it will have an effect on their future job alternatives, and the way they imagine it has affected their job alternatives up till now. It’s the foundation of my research in progress, the outcomes of which haven’t been peer-reviewed.
The findings present that 4 years after ChatGPT’s debut in November 2022, AI has not but resulted in substantial job losses or positive aspects for these employees, opposite to what some AI pessimists and optimists might need predicted.
Solely about 3% of employees stated that they had misplaced a job as a consequence of AI since 2023, whereas roughly 6% indicated they landed a job that didn’t exist earlier than AI. Roughly 9% stated they’ve earned a promotion or development associated to AI.
For AI pessimists like me, it’s shocking that so few of the surveyed employees stated that they had misplaced jobs as a consequence of AI. And it’d elevate a couple of eyebrows to see how unusual it was for employees to point any AI-related job adjustments since 2023.
Restricted influence thus far
YouGov surveyed a pattern of employed U.S. adults meant to mirror the nation’s employed inhabitants by way of age, gender, race, schooling, and different traits.
My 25-question survey included the next three questions regarding AI’s position in job adjustments since 2023:
1) Since 2023, have you ever misplaced a job as a consequence of AI, akin to AI’s automation of your earlier job’s duties?
2) Since 2023, have you ever landed a brand new job that didn’t beforehand exist earlier than the adoption of synthetic intelligence (AI), akin to an AI knowledge labeler or AI compliance specialist?
3) Since 2023, have you ever acquired a promotion or development that you just imagine was associated to AI, akin to your AI expertise?
Roughly 95% of surveyed employees stated no to the primary query; 90% stated no to the second; and 88% stated no to the final. Between 3% and 4% had been “undecided” they skilled any of those AI-related job adjustments. The numbers for every query could add as much as greater than 100% as a consequence of rounding.
What this implies
These outcomes sq. with what some other researchers in the U.S. and elsewhere have found: AI shouldn’t be having widespread results on the labor market at present. On the similar time, job changes in the U.S. are a rarity, so the small percentages of surveyed employees who report AI-related job adjustments have to be seen on this basic context.
An important means AI seems to be affecting job alternatives thus far, in accordance with employees in my survey, is by affording a few of them probabilities for promotion and development. These employees’ AI expertise could very properly be serving to them transfer up the profession ladder, however it’s nonetheless too early to know whether or not any such benefit will final.
Agentic AI, a more moderen type of AI that may act autonomously and automate jobs to a larger extent, could render some present AI experience and expertise much less precious sooner or later.
Surveys like this one supply a snapshot of the reported experiences of a bunch of individuals at one time limit. It’s tough for employees to know for certain whether or not AI is guilty or credit score for his or her job adjustments. Additionally, the survey started in July 2026 when the official unemployment rate was 4.1%, however the workforce was shrinking and AI was reportedly a leading reason for many layoffs.
Those that took the survey are like common U.S. employees in vital methods, however that doesn’t essentially imply their AI experiences and attitudes are typical. They selected to take part in YouGov’s on-line panel, which implies their responses could differ from randomly chosen employees. And none had been jobless, whether or not as a consequence of AI or another excuse.
That stated, my research strikes past many predictions made by economists and analyses produced by AI companies of this know-how’s influence on the labor market by asking employees themselves how they’re experiencing AI. And unlike some research and an AI unemployment tracker in California, my research doesn’t focus simply on job loss; it additionally examines job positive aspects and promotions.
There are remaining questions on how employees are experiencing AI. For instance, who’s getting promoted? Are employees utilizing AI on the job in ways in which could advance their careers? And the way do employees suppose AI will have an effect on their alternatives sooner or later? These are among the many different questions that my survey knowledge can converse to and that I’m presently working to reply.
Jeffrey C. Dixon is a professor of sociology on the College of the Holy Cross.
This text is republished from The Conversation underneath a Artistic Commons license. Learn the original article.
