67.8% of Workers Now Use AI Weekly — But 56.4% Get No Time to Learn It

Nearly 70% of employees now use AI tools beyond ChatGPT at least a couple of days a week — a jump from 39.9% last year — yet 56.4% say their employer allocates no time during work hours for upskilling. Those figures come from Workera's 2026 State of Skills Intelligence Report, published Sept. 23, and they capture a widening gap between AI adoption inside enterprises and the training infrastructure supposed to support it.

The stakes are straightforward. Workera found that 80% of companies feel they are on track for an AI-enabled future in 2026, up from 67% last year, but the people actually operating those AI tools report being squeezed. The survey polled 1,000 salaried professionals at US organizations with 5,000 or more employees, conducted in July 2026 using Pollfish, with year-over-year comparisons drawn from a similar survey in March 2025.

Use is up, learning time is not

More than two-thirds of employees (67.8%) now use AI tools beyond ChatGPT regularly, a nearly 30-point year-over-year increase in adoption. But over 80% of employees (84.3%) spend five hours or less per week on skill development, and 42.5% cite a lack of relevant learning materials as a major obstacle.

Workera founder and CEO Kian Katanforoosh said employees are feeling the strain. "People are feeling like they're cramped; they don't have the time to learn. On top of that, the pace of innovation has kept accelerating," he told ZDNET.

He also noted a shift in what employees are being asked to do. "Last year was more about adoption and access, and now we're past that, and so it's about outcomes and actions, which creates additional stress," Katanforoosh said.

There has been some progress. Nearly six in 10 employees said they've been offered AI-specific training opportunities in the last 12 months, where last year most had not. But organizations are still failing to provide the time and resources for employees to apply what they learn.

Workers are filling the gap themselves

Just under half of employees (46.5%) said they've used tools not provided by their employer for skill development. Of those, 74.6% turned to mainstream AI tools including ChatGPT, Claude, and Gemini — a signal that shadow AI usage persists wherever formal programs fall short.

Katanforoosh's advice to leaders is direct: carve out the time. "I would advise any leader right now to think about giving [their] people time to upskill. Don't think that they will just figure it out without you actually carving out time. With time also comes the psychological safety; give them the psychological safety to experiment [with AI]," he said.

He argued organizations should first define what "AI-ready" means for them, then let employees measure themselves against that standard. "It turns out if you have a standard and you have a measurement, and then a layer of incentives … they push people to accelerate their own learning velocity," Katanforoosh said.

Employees still bet on humans

Confidence in human capability remains high. 76.6% of respondents believed they can do their job better than AI, and only 9.5% said an AI agent could do more than half of their job effectively today. On assessment specifically, just 7% said AI can already evaluate skills better than humans, and 38.5% said AI will never surpass humans at it.

Workera is betting that continuous, opt-in measurement can bridge the trust gap. Its AI-native skills assessment tool Ambient, currently in pilot with over 10,000 signups on its waitlist, measures skills in the flow of work. 41.1% of respondents said they'd opt into such a tool if they owned the data and controlled what's shared; another 31.2% were undecided.

"If you manage to merge work and learning, you're effectively helping someone learn exponentially more than in the past. Today, with AI, we're able to bring learning closer to work because AI understands unstructured data," Katanforoosh said.

The report closes on a point employers ignore at their own cost: building skills is one of the most affordable ways to grow, yet most companies have no way to tell whether their methods work. Katanforoosh tied poor AI training directly to output quality — "the amount of [AI] slop is probably correlated with people knowing or not knowing how to use AI" — meaning enterprises that skip structured upskilling will likely see the results in the work itself.