Career & Negotiation

Heavy AI Users Report More Promotions, but the Survey Can’t Say Why

By Jo Brennan / November 5, 2026 / Updated November 8, 2026

Smiling bearded man in a navy blazer and open white shirt

Numbers first. In a Google-backed survey run by Ipsos, 12% of “AI fluent” workers said AI use helped them get a promotion, against 3% of “AI explorers.” Fluent workers are only 5% of employees in the poll. The gap is real in the sense that people reported it, but the survey cannot tell you that heavy AI use caused any promotion. The answers are self-reported, the sponsor sells AI products, and the fluent group is small.

Key takeaways

Ipsos polled 4,464 employed U.S. adults in December 2025; 40% said they use AI at work, and 5% met the “AI fluent” definition.
Fluent users versus explorers: promotion 12% versus 3%, higher wages 18% versus 4%, better job security 26% versus 7%. All are self-reported.
Put the other way, roughly 88% of fluent users did not report an AI-linked promotion.
The comparison is fluent users against occasional users, not against people who do not use AI.
Treat it as a reason to try AI on real work, not as a promotion strategy.

Who ran the poll, and how

Ipsos fielded the survey and published the results on 19 February 2026. The Ipsos page calls it Google’s AI Works for America poll; Fortune, which got the report as an exclusive, says Google commissioned it with Ipsos. The employee sample was 4,464 employed adults aged 18 to 75, surveyed between 5 and 21 December 2025 on Ipsos’s probability-based KnowledgePanel, in English and Spanish. Results were weighted by gender, age, race and ethnicity, region, metro status, education, income and language to match 2025 Current Population Survey benchmarks. The margin of error for the full employee sample is plus or minus 1.3 percentage points, and Ipsos says subgroup margins are higher. A separate sample of 1,511 hiring managers was surveyed in December too.

Ipsos defines “AI fluent” as employees who redesign their workflows around AI, or who weave it into their work and use it at least weekly across eight or more kinds of tasks. “AI explorers” use it more ad hoc.

What it found

Self-reported measure	AI fluent	AI explorers
Say AI makes them more productive	91%	52%
Median hours saved per week	8	3
Say AI helped them get a promotion	12%	3%
Say AI led to higher wages	18%	4%
Say AI improved job security	26%	7%

The headline ratios in the news (4 times as likely to report a promotion, 4.5 times as likely to report higher wages) are simple division of these percentages, so there is no extra finding behind them. Also note the absolute numbers: even among the most engaged group, only about one in eight credited AI with a promotion.

Elsewhere in the poll, use was highest among educators (64%), office workers (52%) and employees at large businesses (46%). Fortune also reports that 14% of workers were offered AI training by their employer in the past 12 months, and that 53% of non-users said AI does not apply to their work.

Where things stand as of 8 November 2026

I found no follow-up wave or independent replication of the fluent-versus-explorer comparison in the sources I opened. The February numbers are the numbers.

For outside context, Pew Research Center surveyed 5,273 employed U.S. adults in October 2024 and published in February 2025. It found 16% said at least some of their work is done with AI, and another 25% said some of their work could be done with AI but they do not use it much. About one in ten used AI chatbots at work every day or a few times a week. The surveys differ in wording and timing, so do not stack the 16% against the 40%. Both agree on one thing: heavy use is a minority habit.

On pay, the Bureau of Labor Statistics (BLS) reported on 30 October 2026 that wages and salaries rose 3.1% over the 12 months to September 2026. That is the broad yardstick for any claim that AI skills lifted someone’s pay.

What the survey cannot show

Cause. Workers who use AI heavily may already be more ambitious, better placed or employed where AI is encouraged. Promotions may drive access to AI tools rather than the reverse. The poll cannot untangle these.
Self-reporting. These are what workers say AI did for them. Fortune describes the promotion figure as one workers attribute to their own AI skills. Nobody checked payroll or HR records.
The wrong comparison group. Fluent users are compared with explorers, who also use AI. The page I read gives no outcomes for non-users, so it cannot say whether not using AI costs you anything.
Small subgroup. Five percent of 4,464 is on the order of 220 people (my arithmetic; Ipsos does not state the count on the page I read, and weighting changes it). The plus or minus 1.3 points applies to the whole sample, not to this group.
The sponsor. Google makes AI products. Trade coverage dated 25 February described Google pairing the research with a new AI Professional Certificate. A reputable pollster ran it, and the methodology is disclosed. But the topic and the framing were the sponsor’s choice.
Question wording. The full Ipsos PDF could not be read in my tools, so I cannot confirm the exact wording or the base of each percentage.

How to use it without overstating it

The modest takeaway: the people who got most into AI reported more benefits, and the poll cannot say how much of that is the AI. A sensible response is to test it on your own work and keep score.

Pick one task you repeat every week, such as status reports, meeting notes or first drafts.
Time it by hand for two weeks, then time it with AI help. Your own hours saved beat a poll’s median of 8.
Check your employer’s rules on AI tools and confidential data before pasting anything in.
Check the output yourself. You are still responsible for the work.
If it saves real time, tell your manager what it saved and what you did with the time, rather than just naming a tool.
Ask whether your employer offers AI training. Only 14% of workers in the poll had been offered it in the past year.
Do not count on a promotion or raise as a result.

Frequently asked questions

Does using AI heavily get you promoted?

This poll cannot say. Twelve percent of fluent users said AI helped them get a promotion, and 88% did not.

How many workers are AI fluent?

Five percent of U.S. employees in the December 2025 poll, versus 35% explorers. Fluency was defined by Ipsos by how often and widely people use AI, not by a skills test.

Can a Google-sponsored survey be trusted?

The method is solid: a probability-based panel, weighting and a published margin of error. The weaker points are the self-reported outcomes, the small fluent subgroup and the sponsor’s commercial interest. Read it as one data point, not proof.

Should I worry if I do not use AI at work?

These sources give no evidence that not using AI has cost workers promotions. Fortune reports 53% of non-users said AI does not apply to their work. Pew found 52% of workers worried about future AI use in the workplace in 2024, and mixed views on what it means for their jobs.

Before you decide

If you are deciding how much time to put into AI tools, the poll supports trying them on real tasks and measuring the result. It does not support spending money on a certificate or reorganizing your career around a 12% figure. Training, your role and your manager matter more than any poll. For questions about pay, promotion or employment terms, talk to your manager, HR or, for legal questions, an employment attorney.