Nearly Half of US Workers Rely on Free AI Plans at Work

Nearly half are working with the "free version"

In a survey conducted last July by Epoch AI together with Ipsos, involving 1,103 US adult workers, nearly half of the 469 respondents who had used AI for work in the past 7 days relied solely on free plans. Among all respondents, about 47% used free plans, while only around 20% had subscriptions fully covered by their employer. Even adding personal paid subscriptions (about 10%) and mixed employer/personal payments (about 6%), subscriptions involving employer support in any form still fell short of one-third of the total.

Epoch AI summarized: "For most US workers, the default for work-related AI is whatever the free tier provides." Breaking the numbers down by occupation makes this gap even more pronounced.

Science and tech jobs live in a different world

Among computer, engineering, and science occupations (sample of 81), about 52% had subscriptions fully covered by employers, rising to 67% when combined with mixed employer/personal payments. Free plan usage, meanwhile, was only about 20% — less than half the overall average. This suggests that for roles like developers or researchers who use AI tools directly for coding or experimentation, employer-paid subscriptions have effectively become the standard.

By contrast, in management, business, and arts occupations (sample of 207, excluding computer/engineering/science), the employer-support figure dropped sharply to 41%. Free plan usage, at about 43%, roughly matched employer support. This indicates that even among similarly classified "office workers," the extent to which employers open their wallets varies depending on whether the role involves development or research work.

Reliance on free plans increased further in sales/clerical occupations (sample of 72), service occupations (sample of 53), and trade/production/transportation occupations (sample of 49). In trade/production/transportation roles, free plan usage reached about 62%, while employer support was only about 5%.

Subscription patterns by occupation

Occupation	Sample (n)	Free Plan	Employer Support (incl. mixed)
Computer, Engineering, Science	81	20%	67%
Management, Business, Arts	207	43%	41%
Service	53	55%	21%
Sales, Clerical	72	58%	18%
Trade, Production, Transportation	49	62%	8%

The employer support figures in the table combine "fully employer-paid" and "mixed employer/personal" responses. The confidence interval is 90%, and the sample was weighted to be representative of the US wage-earning population aged 18 and over.

Context: an AI price war

These findings coincide with recent steep price cuts by OpenAI and Anthropic on their flagship models. As previously reported by METAL LAB, OpenAI cut both input and output token prices for GPT-5.6 Luna by 80% each, while Anthropic priced Claude Opus 5 at half the cost of its top-tier model, Fable 5, in response to competitive pressure from Chinese models. While falling token prices could, in theory, lower the bar for companies to offer paid subscriptions, this survey reveals the opposite trend. In most occupations, employers still aren't opening their wallets, and workers are largely fending for themselves within free tiers.

Epoch AI noted that the survey followed the BLS's top-level categories under the 2018 Standard Occupational Classification (SOC), and that given the relatively small sample sizes for each occupation group (49 to 207), more weight should be placed on the direction of gaps between occupations rather than on individual figures.

Editor's Take

What this data tells us is simple — the perception that employer-paid AI subscriptions are primarily "developer tools" still runs deep. The 67% figure for computer, engineering, and science occupations suggests AI subscriptions have already been folded into IT budgets much like software licenses. Meanwhile, the roughly 20% figure for other occupations suggests that, from a company's perspective, AI is still classified as a "personal productivity tool." In effect, within the same company, engineering teams may be charging AI subscriptions to a corporate card while marketing teams pay out of pocket.

Anyone who's experienced this gap firsthand won't find these numbers surprising. Whether at a startup or a large corporation, it's common for approval to come easily when engineering teams request AI subscriptions for coding, while marketing, planning, or HR teams requesting the same budget are often told to "use it on your own dime." The reason is simple — development tools can easily prove their value through productivity metrics (code output, bug-fix speed), while other roles struggle to quantify the difference AI usage makes.

For HR and IT managers at domestic companies, this survey offers a useful benchmark. It's worth checking whether the gap in AI subscription support between development and non-development teams at your own company is this wide. If a department shows free-plan reliance above 50%, it's often not because employees aren't using AI, but because the company hasn't paid for it. Given that token prices continue to fall, the cost of rolling out paid subscriptions company-wide is now far lower than it was a year or two ago — leaving less reason to delay a budget review.

What happens over the next few months seems predictable. As the price war continues, companies are likely to expand company-wide licenses regardless of department, eventually rendering this survey's central question — who pays for AI — largely moot. That said, this shift won't happen within this quarter — budget cycles simply don't move that fast.