AI isn’t just changing work. It might be changing workers.
October 15, 2025
Frequent use of the technology is correlated with differences in how people think and feel about their jobs.
Artificial intelligence is changing how many people work. In the process, it also might be affecting how they think and feel about their jobs, their colleagues, and their employers.
Unlike calculators, the typewriter, and software, AI is less a workplace tool and more a workplace teammate. People don’t need inputs to make AI function; they simply have conversations with it. AI is moving workers from Boolean logic to chats, from task checklists to creative thinking.
Since October 2024, ADP Research has asked workers each month about their on-the-job use of generative AI.
Forty-three percent of our survey respondents said they use AI frequently while on the job; 42 percent said they used it rarely or not at all. Frequent users were most likely to be men, and they were most likely to work in technology or information services.
Workplace AI use matters because it has the potential to affect how people think and feel about their jobs. Data from more than 30,000 U.S. respondents to our survey reveals that people who use AI on a daily or near-daily basis report the highest levels of engagement, motivation, and commitment to their work.
These same people, however, also describe weaker connections to their colleagues, report lower productivity, and are more at risk of quitting their jobs.1The ADP Research monthly Worker Sentiment Survey collects data from a stratified, random panel sample of 2,500 workers in the United States. Since October 2024, ADP Research has been collecting monthly data on how frequently people use generative AI to get their work done, to investigate how the frequency of use relates to experiences in the workplace. Data from more than 30,000 respondents was used in this study.
When taken together, the implications of these two findings seem counterintuitive. How can AI usage have the potential to increase engagement and decrease productivity?
Let’s look at how the numbers break down.
Motivation and commitment: Frequent users of AI technology were nearly 3 times more likely to be fully engaged at work, and 3.4 times more likely than less-frequent users to be motivated and committed to their work. Daily users were 5.4 times more likely to say they believe AI will positively impact their job in the next year.
It could be that these heavy users enjoy incorporating the technology into their workday and are accomplishing things they couldn’t do without it. This group also just might be excited about AI’s possibilities.
Connection: While frequent users of AI were more likely to feel fully engaged and committed to their jobs, they were the least likely of any group of users to say they feel strongly connected to their employers and colleagues. Feelings of connection rise significantly among workers who use AI less frequently.
ADP Research defines connection as feeling seen, heard, and valued by others in the workplace.2The ADP Research Connection XPerience Score measures a person’s feelings of being seen, heard, and valued in the workplace based on a study of more than 12,000 survey respondents in the United States. Our methodology and definition for connection can be found at https://www.adpri.org/research/dei-study/.
Engagement and connection: reconciling the contradictions
When it comes to workplace AI use, the disconnect between employee engagement and connection could have myriad sources, including industry differences, occupational roles, and even gender.
People in industries with a high concentration of AI users, such as technology and information, do their jobs differently than people in industries where usage is less common. Workers in customer-facing, team-based roles such as health care or hospitality, for example, might use AI less frequently, but because of the nature of their work they might report stronger connections with colleagues.
In contrast, knowledge workers who are likely to do their jobs independently and out of sight of end users might tend to feel less connected to their colleagues. They also have little visibility into their work outcomes, which can lead to feeling less productive.
Gender, too, might factor into our survey results. Men of all ages tend to be less connected to their peers than women regardless of AI usage.
Productivity: We expected heavy users of AI to report that the technology enhanced their productivity. We found the opposite. The share of workers who consider themselves highly productive was much smaller among those who use AI on a regular basis compared to those who rarely or never use it.
Given generative AI’s nascency, workers probably still are experimenting with it and how to integrate AI into their workflow. As AI takes on repetitive work, some people might find their checklist of tasks growing shorter. Without a box to check, it might be more difficult for these workers to quantify their own productivity.
A lack of formal AI training or development opportunities might be forcing users to take a self-guided approach to acquiring needed skills. They might feel they haven’t mastered the technology’s capabilities.
It’s also possible that frequent users have less confidence in their own productivity based on attributes tied to occupation. By working with AI, these people might feel less ownership of their work or have less certainty in its outcomes.
Intent to leave: Thirty percent of workers who use AI daily said they are actively looking or interviewing for a new job. This share drops steadily as the frequency of AI use decreases, falling to 13 percent among workers who have never tried the technology.
Because heavy adopters of AI are concentrated in the technology and information sectors, intent to leave might be driven by sector dynamics rather than AI usage. But our data did show that even within these sectors, higher AI usage correlated with a higher intent to leave.
It’s unlikely that pay is a factor. The share of workers who feel fairly paid grows steadily as the frequency of AI use increases. This could suggest that AI is augmenting work in a meaningful way.
Alternatively, a lack of connection and the desire to find a job with more human interaction might be driving these responses. And frequent AI users might have helped establish the use of the technology at their organization and are looking for new opportunities to deploy their skill sets.
There’s also a possibility that frequent AI users have a heightened fear of losing their job. Thirty-one percent of workers who use AI daily strongly agree with the statement, “I am scared that my job will be replaced by AI.”
The takeaway
As artificial intelligence becomes more integral to the workplace, it is bringing an interesting combination of enhancements and drawbacks for the people who use it the most.
Considering that the most frequent users of AI, including young men working in tech sectors, are the most likely to say they’re looking to leave their current jobs, employers who rely heavily on these skill sets could benefit from giving extra attention to retention.
Balancing AI users’ engagement and motivation with their tendency to work independently would help employers reap the benefits of AI adoption while mitigating potential costs.
Managers and team leaders could talk with employees about how to improve their connection to others in the organization. They should be alert to the need for training and open to opportunities to use AI. And they could work on alleviating any unfounded fears employees might have about their jobs being at risk.
As in any workplace, employers benefit when their employees know they’re valued and appreciated.
AI users: a demographic snapshot
More than half of workers we surveyed use AI on a regular cadence, and 20 percent called it an essential component of their jobs, something they use nearly every day. But 42 percent use it rarely or never at all. Among those, 21 percent of respondents said they had never even tried AI.3Workers were asked, “How would you characterize your personal use of generative AI?” and told to choose one of the following options: I use it nearly every day, it is an essential part of how I work; I use it multiple times per week to help get work done; I use it multiple times per month; I have tried it but do not use it regularly; I have not tried it yet.
By age and gender
Men were more likely than women to report using AI on the job. Men younger than 40 were the largest group of frequent users, with 29 percent saying they turn to AI in their work nearly every day.
By sector
Workers in people-facing fields tend to use AI less frequently. Educators, health care professionals, public administrators, and hospitality employees—groups that are in regular contact with clients, customers, and the public—were the least likely of all workers to say they use the technology.
At the other end of the spectrum were technology workers, 47 percent of whom said they use AI nearly every day. In information, that share was 41 percent. Only 7 percent of respondents from each of these sectors said they had never tried the technology.
By work type
Knowledge workers were by far the most likely to report using AI on the job; 63 percent said they use the technology at least multiple times a week, and 10 percent said they’ve never tried it. Compare that to repetitive task workers, where only 18 percent use AI at least multiple times a week, and 43 percent have never tried it.4Respondents to our monthly Worker Sentiment Survey are asked to describe their type of work by choosing from three options: I have a level of freedom to use my expertise to create something new. (Our definition of knowledge worker.) I use a level of expertise to solve similar problems each day. (Our definition of skilled task worker.) I do similar repetitive tasks each day. (Our definition of repetitive task worker.)
