AI Mandates at Work Statistics and Company Policy Tracker

Nobody at your company has said you have to use AI. There’s no written policy, no mandatory training, no one checking your usage. And yet, somehow, you feel like you should be using it. Maybe your manager mentioned it in passing. Maybe there’s some unspoken pressure.

If that sounds familiar, you’re in the majority. Only 6% of U.S. workers say AI is formally required at their job, but 48% of regular AI users say they use AI at least partly because their employer expects it, and 21% of all workers say they feel pressured to use it.

Meanwhile, Shopify, Meta, Coinbase, Accenture, Duolingo, and Axon have all made public moves to require, track, or incentivize AI use. Some have reversed course, some have doubled down, and some are still figuring it out.

This page pulls together the most reliable and up-to-date statistics on AI mandates, policies, enforcement, and worker sentiment, along with a dated tracker of specific company examples.

This page pulls together the most reliable 2025 and 2026 statistics on AI mandates, policies, enforcement, and worker sentiment, alongside a dated tracker of named company examples. It’s updated monthly.

How Common Are AI Mandates

Outright “use AI or get fired” rules are still rare, but expectations that look a lot like mandates are much more common.

Among workers who use AI at work a few times a month or more, 7% say they use it only because their employer requires it, another 7% mostly because it’s required, and 34% because of a mix of personal choice and requirement. Added together, that’s 48% of regular users feeling some degree of employer requirement. The survey of 1,533 adults in the U.S. labor force was conducted by Ipsos in June 2026, with sponsorship from the Groundwork Collaborative.

When you ask the broader workforce whether AI is formally required, the numbers drop. Both the CNBC and SurveyMonkey Q3 2026 Quarterly AI & Jobs Survey and the Founder Reports AI in the Workplace Survey land on 6% of workers saying AI use is required by their company.

Pressure, though, is a different measurement. The American Psychological Association’s 2025 Work in America Survey found that 21% of employed U.S. adults feel pressured by their employer to use AI, and 30% worry they’ll fall behind coworkers if they don’t use it. The Harris Poll fieldwork covered 2,017 employed adults in March and April 2025.

The pressure isn’t felt evenly. The APA data shows middle managers feel it most, at 27%, with upper managers at 25%, frontline workers at 15%, and individual contributors at 14%. Pressure is the strongest in the middle, where people are expected to produce more output while also pushing adoption downward.

(Sources: Ipsos, SurveyMonkey, Founder Reports, and APA)

How Many Companies Have Formal AI Policies?

When you ask employers if they have an AI policy, most say yes. When you ask workers, most say no.

Littler’s 2026 Annual Employer Survey of 303 legal and HR professionals found that 68% of employers report a formal AI use policy, up sharply from 2025 (38%). Express Employment Professionals and The Harris Poll ran parallel surveys of 1,002 hiring decision-makers and 1,003 job seekers in November 2025, and found that 78% of hiring managers say their company has AI use policies.

Workers see things differently. Beyond the 55% CNBC/SurveyMonkey figure, the Founder Reports 2026 AI in the Workplace Report found that 44% of U.S. workers say their employer has no clear AI policy or they aren’t sure one exists, based on a survey of 2,078 employed U.S. adults. EisnerAmper’s August 2025 survey, conducted by YouGov among 1,017 U.S. desk workers who use AI, found that just 36% say their company has a formal AI policy.

The disconnect is even wider outside the U.S. The KPMG and University of Melbourne Trust, Attitudes and Use of AI Global Study, covering over 48,000 people in 47 countries, found that only 40% of employees say their workplace has any policy or guidance on generative AI use.

Company size matters a lot here. Our survey found 59% of workers at companies with fewer than 10 employees have no clear AI policy or aren’t sure, compared with 34% at companies with 1,000 or more employees.

The simplest explanation for the employer-worker split is that policies are being written at the top and not communicated down. Gallup’s quarterly tracking supports this. Only 26% of U.S. employees said in August 2026 that their organization had communicated a clear AI plan, up from 22% in May 2025. Written policies move faster than internal communication.

(Sources: Littler, Express Employment Professionals, Founder Reports, EisnerAmper, University of Melbourne, and Gallup)

How Companies Enforce AI Use

The 2026 AI Adoption in the Enterprise survey from WRITER and Workplace Intelligence, which polled 1,200 C-suite executives and 1,200 employees across the U.S. and Europe between December 2025 and January 2026, found that 60% of executives plan to lay off employees who can’t or won’t use AI. Another 77% say employees who refuse to become AI-proficient won’t be considered for promotions or leadership roles.

Those numbers are significant, but the same survey also provides some interesting counterweight. 75% of those executives also admit their company’s AI strategy is “more for show than actual internal guidance.”

Hiring is becoming a backdoor mandate. KPMG’s American Worker Survey of more than 2,100 employees at large U.S. organizations found that 76% of managers actively look for AI-capable talent when hiring. For job candidates, that’s essentially a mandate to have experience before getting the job.

Government is in the mix too. OMB Memorandum M-25-21, signed in April 2025, directs federal agencies to appoint Chief AI Officers and “establish clear expectations for their workforce on appropriate AI use.” CFO Act agencies had 180 days to publish AI strategies on their websites.

(Sources: Writer, KPMG, and Whitehouse.gov)

Company AI Mandate Tracker

These companies have required, tracked, or walked back AI use (listed with the most recent first).

Company	Policy Type	What Happened	Date	Source
Axon	Productivity target	About 2,200 R&D employees were given a mandatory 10% productivity-gain target. Teams exceeding 15% receive stock grants.	Oct 2026	Fortune
Amazon	Usage tracking (reported)	Employees reportedly used an internal AI tool on trivial tasks to climb a usage leaderboard. Amazon reportedly said this wouldn’t factor into reviews.	May 2026	Fortune, citing FT
Duolingo	Reversed	CEO Luis von Ahn said the company “backtracked” on evaluating AI use in performance reviews.	Apr 2026	Business Insider
Meta	Leaderboard removed	An employee-built “Claudeonomics” leaderboard ranked the top 250 token users. Usage topped 60 trillion tokens in 30 days. It was shut down after data leaked externally.	Apr 2026	Fortune
Accenture	Promotions	Weekly AI tool log-ins of some senior staff are monitored. “Use of our key tools will be a visible input to talent discussions.”	Feb 2026	Fortune
Meta	Performance reviews and pay	“AI-driven impact” became a core performance expectation starting in 2026. Checkpoint tool counts AI-generated code and tracks over 200 data points. Top-rated employees eligible for a 200% bonus multiplier.	Nov 2025–Feb 2026	HR Grapevine
Accenture	Workforce exits	CEO Julie Sweet: “We are exiting on a compressed timeline, people where reskilling, based on our experience, is not a viable path.” Over 550,000 employees trained in genAI fundamentals. About $865M in business optimization costs.	Sep 2025	Accenture earnings call transcript
Coinbase	Required, with firings	CEO Brian Armstrong gave engineers about a week to onboard Copilot or Cursor. Engineers who hadn’t onboarded were called to a Saturday meeting, and some “got fired.”	Aug 2025	Fortune
Amazon	Encouraged	CEO Andy Jassy memo urged employees to learn AI and said genAI will “reduce our total corporate workforce” in the coming years.	Jun 2025	About Amazon
Microsoft (Developer Division)	Required	Memo from Julia Liuson said “using AI is no longer optional, it’s core to every role and every level.” Managers asked to consider AI use when evaluating employees.	Jun 2025	Business Insider
Duolingo	Required (original)	“AI-first” memo factored AI into hiring and performance reviews. Headcount added only if work can’t be automated. Contractors phased out for AI-capable work.	Apr 2025	Computing
Shopify	Required	“Reflexive AI usage is now a baseline expectation at Shopify.” Teams must demonstrate why AI can’t do the work before requesting more headcount.	Apr 2025	Digital Commerce 360
U.S. federal government	Directive	OMB M-25-21 requires agencies to appoint Chief AI Officers and establish workforce expectations on appropriate AI use. CFO Act agencies had 180 days to publish AI strategies.	Apr 2025	OMB Memorandum M-25-21

Tracking, Dashboards, and Tokenmaxxing

Formal monitoring of employee AI use remains uncommon overall. Only 22% of AI-using U.S. desk workers say their employer actively monitors their AI usage, based on EisnerAmper’s 2025 YouGov-fielded survey of 1,017 workers. Another 41% say they inform their manager or seek permission before using AI.

At big tech companies, measurement has arrived fast, and it hasn’t always gone smoothly. Meta’s “Claudeonomics” leaderboard was built by an employee to rank the top 250 token users across the company’s more than 85,000 employees. According to Fortune’s coverage, employee token usage topped 60 trillion over a 30-day window, and the single highest-ranked user averaged 281 billion tokens. The dashboard was shut down two days after the news broke, with an internal note saying it was shuttered because data had been shared externally.

Amazon has reportedly faced the same gaming problem. Fortune reported that Amazon employees ran the company’s internal AI tool on trivial tasks to inflate their token counts. Amazon reportedly told employees that tokenmaxxing wouldn’t factor into reviews, though several told reporters they suspected managers monitored it anyway.

The measurement problem is a predictable one. Once usage becomes visible on a leaderboard, employees optimize for the leaderboard. Token volumes don’t measure productivity, output quality, or business value. They measure how much AI was used.

(Sources: EisnerAmper, Fortune, and Fortune)

Shadow AI and Rule Breaking

When companies push AI without providing the right tools or guidance, workers find their own.

The KPMG and University of Melbourne global study found that 57% of employees hide their use of AI at work and present AI-generated work as their own. Almost half admit to using AI in ways that break company policies. The survey covered more than 48,000 people in 47 countries, fielded between November 2024 and January 2025.

EisnerAmper found that 28% of AI-using desk workers would use AI at work even if it were banned.

The WRITER survey went further. 29% of employees admit to sabotaging their company’s AI strategy, including by using unapproved tools or refusing to use AI. Among Gen Z employees, that figure rises to 44%.

Shadow AI is a real issue. IBM’s 2025 Cost of a Data Breach Report, conducted by the Ponemon Institute across 600 organizations globally, found that 63% of breached organizations had no AI governance policy or were still developing one. One in five reported a breach involving shadow AI, and organizations with high levels of shadow AI faced roughly $670,000 in additional breach costs compared with those with low or no shadow AI.

The underlying driver is a lack of alternatives. BCG’s 2025 AI at Work survey of more than 10,600 employees in 11 markets found that over half will find their own AI tools if their employer doesn’t provide them.

(Sources: University of Melbourne, EisnerAmper, WRITER, IBM, and BCG)

AI Training and Support Statistics

Unfortunately, usage is often mandated without adequate training.

KPMG’s survey of large U.S. organizations found that 87% of employees now use AI at least weekly and 51% use it daily, up nearly 20 points from the prior year. Only 42% of organizations make AI training mandatory, and 84% of workers want more training.

Workers and hiring managers rarely agree on much. This is one of the exceptions. Express Employment Professionals and The Harris Poll found that 83% of job seekers say companies need to formally train employees on AI, and 86% of hiring managers say formal AI training should be a company priority. Only 36% of companies, however, provide a list of approved AI tools.

Support changes everything. BCG’s data shows that the share of employees feeling positive about generative AI rises from 15% to 55% with strong leadership support. Yet only a quarter of frontline employees say they receive that support, and only a third say they’ve been properly trained.

Manager behavior also impacts employee AI usage. Gallup’s Q1 2026 analysis of 23,717 employed U.S. adults found that frequent AI use rises to 78% among employees who strongly agree their manager supports it, compared with 44% among those who don’t.

(Sources: KPMG, Express Employment Professionals, BCG, and Gallup)

Do AI Mandates Actually Pay Off?

The business case for mandates assumes AI makes workers more productive. But the evidence isn’t clear.

Founder Reports’ survey of 2,078 U.S. workers revealed at companies that require AI use, 73% of workers have had to fix a coworker’s AI-reliant work, and 17% say it happens regularly. At workplaces with no AI policy, the fix rate drops to 30%. The rework rate scales directly with how much AI a workplace mandates, and that cost lands on the coworkers who have to review and correct output.

A randomized controlled trial from METR, published on arXiv in July 2025, adds another counterweight. Sixteen experienced open-source developers working on 246 real tasks in their own mature codebases took 19% longer to complete tasks when AI tools were allowed, even though the developers believed AI had cut their completion time by 20%. The sample is small, and the setting is specific, so this isn’t a verdict on AI productivity broadly. But it is a caution about the gap between felt productivity and measured productivity.

Section’s AI Proficiency Report, which surveyed over 5,000 knowledge workers at companies with 1,000 or more employees, found that 24% of workers save no time using AI, and 85% have beginner-level AI use cases or none at all.

On the employer side, U.S. Census Bureau researchers found that 66% of AI-adopting firms use AI only to augment worker tasks, not replace them. Most firms (57%) use AI in just one to three business functions. The authoritative government data looks nothing like the “AI will replace everyone” framing of some CEO memos.

(Sources: Founder Reports, arXiv, Businesswire, and US Census Bureau)

How Workers Feel About AI Mandates

Worker sentiment has shifted sharply in a short period of time.

Ipsos’s June 2026 survey found that 67% of U.S. workers and job seekers say AI in the workplace will make the worker experience worse by eliminating jobs and pressuring workers to be more productive, versus 30% who say it will improve work. 51% say the benefits will go only or mostly to business owners and executives.

The anxiety is specific and localized. The Federal Reserve Bank of Philadelphia’s LIFE Survey of 2,632 employed U.S. adults, fielded between December 2025 and January 2026, found that 27% say AI is affecting their own job or career, while 62.9% say it’s affecting the job market overall. Workers see AI changing almost everything.

KPMG’s data shows 52% of employees at large organizations worry AI could eventually displace their jobs, nearly double the prior year’s figure. Six in ten Gen Z workers believe AI could replace their role within two years, compared with three in ten older employees.

Job security worries are higher where AI is being pushed hardest. BCG found that 46% of employees at companies undergoing comprehensive AI-driven workflow redesign worry about job security, compared with 34% at less advanced companies.

(Sources: Ipsos, Federal Reserve Bank of Philadelphia, KPMG, and BCG)

Context on AI Adoption

The mandate push is happening on top of real but uneven AI adoption.

Gallup’s quarterly tracking shows 53% of U.S. employees use AI at work at least a few times a year as of August 2026, with 16% using it daily, up from 8% in May 2025.

The business numbers look different. The U.S. Census Bureau’s Business Trends and Outlook Survey reported that 19.8% of U.S. businesses used AI in any business function during the two weeks ending May 3, 2026. Information (39.7%) and Finance and Insurance (33.9%) were the leading sectors. Firm size drives most of the variation, with 37% of firms with at least 250 employees using AI compared to less than 20% of firms with four or fewer employees.

The workforce is using AI faster than its employers are adopting it, which may be why employee mandates are showing up before organizational infrastructure.

For a fuller look at adoption data, see our AI in the Workplace Statistics for 2026.

(Sources: Gallup and US Census Bureau)

What Employers Should Do Before Mandating AI

The data points to a few practical steps for any company considering an AI mandate. Each one connects to a statistic from above.

• Communicate the policy, don’t just write it. The disconnect exists because policies aren’t reaching employees. A policy no one will do no good.
• Define measurement by outcome, not usage. Meta’s Claudeonomics leaderboard became a gaming target within weeks. Duolingo reversed its AI performance-review policy after employees said it felt performative.
• Train before enforcing. Only 42% of large organizations make AI training mandatory, while 84% of workers want more training. Mandates without training produce the rework problem seen in the Founder Reports AI in the Workplace study.
• Provide approved tools, in writing. A list of approved tools reduces shadow AI.
• Budget for review time. Rework rates jump at AI-required workplaces. Someone has to check the output, and that time has to come from somewhere.

About This Data

This page draws on 2025 and 2026 surveys and primary documents. Where possible, we’ve linked to the original PDF, press release, or government document rather than news coverage.

Vendor-sponsored surveys are labeled as such. different things and should be read side by side rather than as competing claims. The company tracker uses the earliest primary source we could locate for each entry.

Frequently Asked Questions