Only 43% of Workers Use AI at Work. The Other 57% Are Falling Behind Fast.

The labor market has been sorting workers into two groups for a while now. The sorting has just gotten a lot more expensive to ignore.

New data from Indeed’s Hiring Lab, published in January 2026, reveals that only about 43% of U.S. workers reported regularly using AI at work in 2025. Perhaps more striking, roughly 40% described themselves as actively disengaged from AI tools entirely.

Not skeptical. Not cautious. Actively disengaged.

In a job market where AI-related job postings are growing at 134% above pre-pandemic levels while total job postings sit just 6% above that same baseline, the gap between those two groups is starting to show up in wages, hiring rates, and long-term career trajectory.

This is the story of how the AI adoption curve is reshaping the labor market from the inside out, and why the numbers are more consequential than most workers realize.

The Adoption Divide Is Wider Than It Looks

When you hear that AI adoption at work is surging, the headline numbers can be misleading.

Gallup’s Q4 2025 workplace data found that nearly half of U.S. workers still never use AI in their role at all. Frequent use, meaning several times per week or more, clocked in at just 27% for white-collar workers.

The St. Louis Federal Reserve’s Real-Time Population Survey adds more context. As of August 2025, generative AI adoption at work had reached approximately 54.6% for any use whatsoever. But workers report that AI accounts for only about 5.7% of their total work hours, up from 4.1% in late 2024.

The technology is spreading. But shallow use and deep use are very different things.

The breakdown by industry makes the divide even clearer:

• Technology: 77% total AI use, with 57% using it frequently
• Finance: 64% total AI use
• Higher Education: 63% total AI use
• Healthcare and Government: 41-43% total AI use
• Retail: 33% total AI use, the lowest of any tracked sector

The split between leaders and individual contributors is just as stark. Gallup found that frequent AI use among organizational leaders has risen from 17% to 44% since mid-2023. Among individual contributors, it has grown from 9% to 23% over the same period.

The people at the top are integrating AI into their daily work at twice the rate of the people doing frontline work. That gap has real implications for who gets promoted and who gets passed over in the next round of performance reviews.

Interview Guys Take: The adoption divide is not just a technology story. It is an access story. Leaders have more discretionary time, more exposure to training, and more latitude to experiment. The 40% who are actively disengaged are not necessarily resistant. Many of them simply have not been given a real reason or opportunity to engage.

The Training Gap Nobody Wants to Talk About

Here is a number that reframes the entire adoption debate.

According to Randstad research, only 13% of workers have received any AI training from their employers. Meanwhile, 55% of workers say they want more AI training to protect their careers.

The workforce is not disengaged because workers do not care. Many are disengaged because they have been left to figure it out on their own, and figuring it out alone is genuinely hard when you are working full-time and no one around you is doing it either.

This has produced an unusual dynamic. Workplace surveys show that approximately 78% of professionals who do use AI at work bring their own tools, a trend researchers call Bring Your Own AI (BYOAI). Workers are outpacing their employers’ formal adoption strategies, downloading ChatGPT, Claude, and other platforms independently to get more done.

But that organic adoption is concentrated among workers who already have high initiative and some baseline of technical comfort. The 40% who are disengaged are not in that group.

The SurveyMonkey 2025 Workplace AI report adds a cautionary note to even the adoption that is happening. It found that 29% of employees admitted to using AI on the job without telling their manager, and 57% of employees admit they do not check AI output for accuracy.

The adoption that is happening is often unofficial, inconsistent, and undertrained. That is not a strong foundation for the measurable productivity gains employers are hoping for.

What the Wage Data Actually Says

The financial stakes of the adoption divide are not theoretical.

PwC’s 2025 Global AI Jobs Barometer analyzed close to one billion job postings across six continents. Their finding: workers with AI skills command a 56% wage premium over peers in the same occupations without those skills. That figure was 25% just one year earlier.

The premium more than doubled in twelve months.

This is not limited to tech roles. PwC found that wage premiums for AI skills exist in every single industry analyzed, including finance, healthcare, professional services, and manufacturing.

Our own deep dive into what AI skills actually pay found that Lightcast’s analysis of 1.3 billion job postings put the AI skills premium at roughly $18,000 more per year at median salary levels. For workers with two or more specific AI skills, that premium climbs to 43%.

A wage premium that doubled in a single year is not a stable, predictable career investment. It is a signal that the market is still in the early, fast-moving phase of pricing this skill set. Historically, that means early movers capture outsized returns before the premium normalizes as supply catches up.

The hiring composition tells the same story from a different angle. According to Indeed’s data, at the end of 2025, tech job postings that mentioned AI were about 45% higher than pre-pandemic levels, while total tech postings were 34% below pre-pandemic levels. AI mentions are the difference between a growing slice and a shrinking sector. We broke down how that split plays out across industries in our analysis of the best and worst jobs in the AI age.

Interview Guys Take: A 56% wage premium is not a niche advantage for data scientists. It is a general market signal. When PwC finds that premium appearing across every industry they analyzed, including industries with relatively low AI exposure, it becomes very hard to argue that learning AI tools is optional. The workers who frame AI literacy as “that is for tech people” are making a financial decision, whether they know it or not.

The Jobs Numbers Tell the Same Story

The wage gap would be notable on its own. The employment data reinforces it.

Indeed’s Hiring Lab shows a stark bifurcation right now. Total job postings in January 2026 sit just 6% above pre-pandemic baseline. Job postings that mention AI terms have surged to 134% above that same baseline.

The two trend lines are moving in opposite directions. AI-adjacent hiring is expanding in a market where general hiring is essentially flat. We covered the broader implications of that in our piece on AI jobs paying $200K and above.

The skills evolution underneath that hiring shift is equally fast-moving. PwC found that the skills demanded in AI-exposed jobs are changing 66% faster than in non-AI-exposed roles, up from 25% the year before. That rate of change is not slowing. It is accelerating.

For workers who are disengaged from AI now, the learning curve grows steeper every quarter they wait.

The WEF Future of Jobs Report 2025 found that 85% of employers plan to prioritize upskilling their workforces over the 2025-2030 period. But employer intentions and employer actions are different things. The 13% training figure from Randstad suggests most employers are still firmly in the intention phase.

Waiting for your company to train you is not a reliable strategy.

The Productivity Gap Is Now Measurable

One of the bigger questions researchers have been wrestling with is whether AI adoption actually translates into measurable productivity gains. The St. Louis Fed now has a clear answer.

Their analysis suggests that generative AI has increased U.S. labor productivity by approximately 1.3% since ChatGPT’s release. That sounds modest, but it represents a meaningful acceleration above the pre-pandemic trend of 1.43% annual productivity growth. From Q4 2022 through Q2 2025, annualized productivity growth ran at 2.16%, with the excess largely attributable to AI adoption.

The speed of the adoption curve itself is striking. The personal computer achieved 19.7% adoption three years after its mass-market introduction. The internet hit 30.1% at the same three-year mark. Generative AI reached 54.6% adoption in the same timeframe, the fastest technology adoption curve in the modern labor market.

When a technology spreads that fast, the gap between early adopters and laggards widens before most people realize the race has started.

Workers in the actively disengaged 40% are not just missing a skill. They are missing productivity gains that employers can now measure directly, and that measuring is starting to show up in hiring decisions and compensation conversations. Our overview of essential AI skills for workers breaks down which specific capabilities are driving the most practical value.

Interview Guys Take: The productivity story is what makes the adoption gap a structural issue rather than a preference issue. When the Fed can document an economy-wide productivity acceleration attributable to AI, individual employers can run the same calculation at the team level. A manager looking at two otherwise comparable employees, one of whom consistently uses AI to do more in less time, is looking at a real performance difference. Not a trendy resume line.

What the Disengaged 40% Actually Looks Like

The actively disengaged group is not monolithic.

Gallup found that AI use at work is most concentrated in remote-capable knowledge work and least common in production and service roles. Workers in retail, manufacturing, healthcare, and community services are the least likely to have any meaningful exposure to AI tools in their daily work.

Age and organizational level also matter more than most people expect. McKinsey found that 62% of workers aged 35-44 report strong AI skills, while adoption among both younger and older workers is notably lower. That cuts against the common assumption that Gen Z workers are uniformly AI-fluent. They are not.

Gallup’s data adds one more finding worth sitting with: 41% of organizations have not implemented any AI tools at all, and 21% of employees do not even know whether their organization has.

If a significant share of workers genuinely cannot tell whether their company uses AI, the disengagement number starts to make a lot more sense. It is hard to engage with a tool that has never been introduced.

For workers thinking about how to position themselves in this environment, our breakdown of how employers will evaluate AI skills in 2026 is worth reading alongside the salary and adoption data here. The picture that emerges is of a hiring market already selecting for AI fluency, even in roles that would not traditionally be thought of as tech-adjacent.

The Big Picture

The 43% adoption figure from Indeed’s Hiring Lab is a snapshot of a labor market mid-transformation.

The fact that roughly 40% of workers are actively disengaged from AI is not a permanent condition. But the window for that group to close the gap without significant career consequences is narrowing.

The wage data from PwC, the hiring data from Indeed, and the productivity data from the St. Louis Fed all point in the same direction. AI fluency is moving from a differentiating advantage to a baseline expectation, and it is doing so faster than almost any labor market transition in recent history.

The workers and employers who treat the current adoption gap as a temporary anomaly that will sort itself out are likely misreading the trajectory. The gap is not closing on its own. A 56% wage premium that doubled in a single year is not a sign of a market stabilizing. It is a sign of one still accelerating.

For a broader look at how this fits into the larger reshaping of work, our comprehensive report on the state of AI in the workplace covers the full scope of what the research is showing.

---