Employees Fake AI Use: 48% Exaggerate to Bosses [2026]

Nearly half of American office workers are lying about how much artificial intelligence they actually use on the job, according to a new report that is spreading fast across tech and HR media this week. The behavior has a name now: performative AI. It describes employees who talk up their AI skills, drop tool names into meetings, and claim AI-assisted wins they didn’t really achieve, all to look competent in front of managers who are themselves under pressure to prove their teams have “gone AI-first.”

The trend was detailed in a TechRadar report published October 12, 2026, and updated the following day, drawing on data from workforce analytics firm Visier. The headline numbers are stark: 48% of employees say they have exaggerated their AI usage or expertise to colleagues or managers “some or most of the time,” and 45% say they feel pressured to use AI at work at all. Those two figures, taken together, paint a picture of a workforce performing enthusiasm it doesn’t necessarily feel, for an audience of bosses who may not be checking too closely.

·

Google · Preferred Sources

Don't miss new tech stories on Google

Add TrendinTech once in the Google app and our stories appear in your news suggestions.

Add Now

What the Visier Report Actually Found

Visier’s research, cited across multiple outlets this week, surveyed full-time employees in the United States about their real versus reported AI habits at work. The core finding is not that AI adoption is low. It’s that reported AI adoption and real AI adoption have become two different numbers, and the gap between them is being papered over by employees who would rather fake competence than admit uncertainty. According to TechSpot’s coverage of the survey, the sample consisted of 1,000 full-time US employees, giving the findings a reasonably broad base for a workplace-culture study of this kind.

TechSpot’s writeup also credits Visier with coining the term “Performative AI” to describe this specific behavior: claiming or implying AI fluency that doesn’t match actual day-to-day use. That’s distinct from general workplace dishonesty. It’s a very particular, AI-era flavor of impression management, where the currency being faked is not effort or hours worked but technical sophistication with a specific category of tool.

The pressure side of the equation matters just as much as the faking side. When 45% of employees say they feel pressured to use AI at work, that pressure has to be coming from somewhere, whether it’s explicit management directives, competitive anxiety about coworkers, or a vaguer sense that the entire corporate world has decided AI fluency is now mandatory. Fast Company’s reporting on the same Visier data, published in late September 2026, frames the exaggeration as a coping mechanism for employees trying to keep up with expectations they can’t always meet on a realistic timeline.

Performative AI Isn’t a Brand-New Phenomenon

What makes this week’s numbers notable is scale, not novelty. Faking AI use at work has been documented in smaller, earlier surveys stretching back over a year. In July 2025, The Register reported on a Howdy.com survey that found roughly one in six US workers, about 16%, admitted to pretending to use AI specifically to please their bosses. That same survey found that roughly three-quarters of employers expected employees to use AI in some form, with about half of those expectations formalized into official policy and the rest left as informal, ambient pressure.

The jump from a 16% “pretend to use AI” figure in mid-2025 to a 48% “exaggerated usage or expertise” figure by October 2026 isn’t a perfectly apples-to-apples comparison, since the two surveys measure slightly different behaviors and come from different firms. But the direction is consistent across every report on this topic: more employees, not fewer, are gaming the appearance of AI fluency as the technology becomes more embedded in performance reviews, team meetings, and day-to-day workplace vocabulary.

A separate data point reinforces the trend line. Reporting from 36Kr in October 2026 cites a Howworks.ai survey finding that 51% of American workers are worried about losing their jobs to AI, while 72% of professionals say they lack confidence in their own ability to adapt to AI tools, and 16% admit to pretending to use AI at work just to please their bosses. Even allowing for methodological differences between firms, the consistent presence of a “job security anxiety” layer underneath every one of these reports suggests performative AI is less about vanity and more about fear.

The Data Across Reports: A Side-by-Side Look

Because several firms have independently measured some version of this behavior over the past 15 months, it’s worth lining the findings up side by side. The table below pulls directly from the named reports and outlets covering performative AI, without blending incompatible numbers into a single misleading average.

Report / Source	Published	Sample	Key Finding
Visier (via TechRadar, TechSpot, Fast Company)	Sept-Oct 2026	1,000 full-time US employees	48% exaggerated AI usage or expertise to colleagues/managers; 45% felt pressured to use AI
Howdy.com (via The Register)	July 2025	US workers	16% pretend to use AI to please bosses; ~75% of employers expect some AI use
Howdy.com (via Newsweek)	August 2025	US professionals	Roughly 16% sometimes pretend to use AI
Click Boarding internal project (via HRZone)	April 2026	Case-study based	Employees fake AI use to comply with internal mandates and protocols
Howworks.ai (via 36Kr)	October 2026	Not specified	51% worried about job loss to AI; 72% lack confidence adapting; 16% pretend to use AI

Read across the rows, the pattern holds even though the exact percentages vary by firm and methodology: a meaningful share of the workforce is managing perception around AI rather than, or in addition to, actually building AI skill. The Visier figures sit at the high end of that range, but they are also the most recent and the most widely corroborated across independent outlets this month.

Why Employees Are Faking It: Fear, FOMO, and Mixed Signals

Every outlet covering this story points to roughly the same underlying driver: employees don’t want to be seen as falling behind on AI, even when the tools themselves are confusing, inconsistent, or poorly rolled out by their employer. TechRadar’s framing is direct about the fear angle, noting that workers are lying about AI adoption partly out of concern that admitting low usage, or admitting confusion about how to use AI tools effectively, could mark them as expendable in a workplace where AI fluency has become a proxy for value.

That fear doesn’t exist in a vacuum. It sits on top of a corporate messaging environment where executives, boards, and public companies have spent much of 2025 and 2026 publicly tying AI adoption to productivity narratives, headcount decisions, and competitive positioning. When leadership frames AI fluency as existential, individual employees respond rationally by performing fluency even when their actual comfort with the tools hasn’t caught up. The HRZone piece on the Click Boarding project describes this directly: employees pretending to use AI tools specifically to comply with internal protocols, not because the tools solved a real problem for them.

There’s also a simpler, more human explanation that doesn’t require any grand theory of corporate AI anxiety: nobody wants to look slow in a meeting. If a colleague casually mentions that an AI assistant drafted their quarterly report in ten minutes, the social cost of admitting “I don’t really use that” can feel higher than the cost of a small exaggeration. Multiply that dynamic across thousands of standups, performance reviews, and Slack threads, and a 48% exaggeration rate stops looking like an outlier and starts looking like simple herd behavior.

The Manager’s Blind Spot

Performative AI only works as a strategy if managers aren’t actually verifying the claims. That’s the uncomfortable part of this story for corporate leadership. A workforce that is 45% pressured and 48% willing to exaggerate is, by definition, a workforce whose self-reported AI metrics cannot be trusted at face value, which creates a serious measurement problem for any company trying to track real AI ROI through usage surveys, internal dashboards, or employee self-assessment during review cycles.

This is where the story stops being a quirky workplace trend piece and starts mattering to anyone making budget decisions about enterprise AI licenses, Copilot seats, or custom LLM tooling. If a meaningful chunk of “active users” reported in an internal adoption dashboard are actually just employees who opened the tool once to avoid looking bad, then license utilization numbers used to justify renewal spend are inflated by an unknown but apparently nontrivial margin.

Market Impact: What This Means for AI Vendors and IT Buyers

For companies selling workplace AI tools, performative AI is a double-edged signal. On one hand, high reported usage, even if partly inflated by employees gaming perception, still shows up as engagement in vendor dashboards and can support renewal conversations. On the other hand, enterprise buyers who read this week’s coverage closely are likely to start asking harder questions about what “usage” actually means in their own seat-utilization reports, especially heading into 2027 budget cycles where every software line item is getting scrutinized against measurable productivity gains.

That scrutiny has a direct effect on how IT and HR leaders structure AI rollouts going forward. Instead of counting logins or seat activations as a proxy for adoption, more organizations are likely to push toward outcome-based metrics, tracking whether AI-assisted work actually changed cycle times, error rates, or output volume rather than trusting self-reported survey data or dashboard login counts. That shift benefits vendors who can prove measurable task-level impact and disadvantages vendors whose value proposition rests mainly on seat counts and usage claims that employees have an incentive to inflate.

It also changes the calculus for AI training budgets. If a large share of reported fluency is performance rather than substance, the actual skills gap inside most companies is wider than internal surveys suggest, which argues for more structured, verifiable AI training programs rather than the ad hoc “figure it out yourself and report back” approach many companies adopted during the first wave of generative AI rollouts in 2023 and 2024.

Pressure Points: A Closer Look at the Sentiment Data

Pulling the pressure-related figures together from the reports covering this story gives a fuller picture of workforce sentiment around AI mandates, separate from the usage-faking numbers themselves.

Metric	Figure	Source
Employees who exaggerated AI usage/expertise “some or most of the time”	48%	Visier, via TechRadar
Employees who felt pressured to use AI at work	45%	Visier, via TechRadar
Workers who admit to pretending to use AI to please bosses (2025 benchmark)	16%	Howdy.com, via The Register
Employers who expect AI use in some capacity	~75%	Howdy.com, via The Register
Workers worried about losing their job to AI (2026)	51%	Howworks.ai, via 36Kr
Professionals lacking confidence adapting to AI tools (2026)	72%	Howworks.ai, via 36Kr

The two numbers that stand out most are the job-security anxiety figure (51%) and the confidence gap (72%). Together they suggest performative AI isn’t really about employees trying to seem impressive for its own sake. It’s about employees trying to avoid being seen as the person who gets cut when a department decides to “lean into AI” during the next round of restructuring.

Historical Context: From “Resume Keyword Stuffing” to “Performative AI”

Workplace exaggeration around hot technology isn’t new; it’s just found a new target. The cloud computing boom of the early 2010s produced its own version of resume inflation, with job seekers and employees alike padding their familiarity with AWS or Azure to match hiring demand. The “big data” era of the mid-2010s saw similar inflation around tools like Hadoop and Spark. What’s different about the performative AI moment is the compressed timeline: AI tools moved from novelty to expected-competency in a little over two years, giving the workforce far less runway to build real skill before faking it became the path of least resistance.

It also matters that this round of hype arrived during a period when layoffs tied explicitly to AI efficiency gains became a recurring headline across the tech sector through 2025 and into 2026. That combination, rapid tool proliferation plus visible job-cut narratives tied to AI, is largely what separates performative AI from earlier waves of resume-padding. The stakes, at least as perceived by employees, are higher this time.

Competition: How Different Industries Are Responding

Coverage of the Visier data hasn’t broken out industry-by-industry figures in detail, but the broader reporting pattern across tech, HR, and business media this week suggests the behavior is being discussed most urgently in knowledge-work sectors, software, marketing, consulting, finance, where AI tools have been pushed hardest and fastest into daily workflows. Sectors with heavier physical or regulated operations, where AI adoption has been slower and more deliberate, appear less frequently in the reporting, likely because the pressure to visibly “use AI” in every task is lower when AI isn’t yet central to the job itself.

HR technology vendors are starting to respond to this gap directly. The HRZone piece on the Click Boarding project frames performative AI adoption as a new category of employee disengagement that traditional engagement surveys don’t capture, since an employee can score well on an “I use AI regularly” survey question while contributing nothing of substance from that usage. That’s prompting some HR teams to explore task-level verification rather than self-report surveys as the basis for measuring adoption, a shift that mirrors the broader move toward outcome-based metrics on the vendor side.

The Productivity Paradox Hiding Inside the Numbers

There’s a deeper irony buried in these findings. Companies have spent two years pushing AI adoption specifically to capture productivity gains, yet a near-majority of employees are now spending energy managing the appearance of adoption rather than the substance of it. That’s not a free action. Every hour spent crafting a plausible AI-usage narrative for a manager, every Slack message mentioning an AI tool that wasn’t actually used for the task at hand, is an hour not spent doing real work or honestly building the skill the company is paying for.

This creates a strange possibility worth naming directly: in some organizations, the performative layer around AI adoption may be actively offsetting whatever genuine productivity gains the tools are producing elsewhere. Until companies find better ways to measure real task-level impact instead of self-reported usage, that offset is largely invisible in the metrics leadership actually sees.

Five Predictions for Workplace AI Culture Through 2027

• Expect more companies to shift from self-reported AI usage surveys to task-level telemetry, measuring what AI tools actually touched in a workflow rather than asking employees to self-certify fluency.
• Expect at least one major HR software vendor to launch or expand a product explicitly marketed around detecting “AI adoption gaps” between claimed and actual usage, building directly on the Visier framing of performative AI.
• Expect performance review language to shift away from vague “AI-forward” mandates toward specific, auditable deliverables, as managers grow wary of unverifiable self-reported AI claims.
• Expect continued divergence between knowledge-work sectors, where performative AI pressure stays high, and operations-heavy sectors, where physical and regulatory constraints keep AI mandates more measured.
• Expect follow-up surveys in 2027 to show the exaggeration rate holding steady or climbing further unless employers change how they measure and reward AI use, since nothing in the current incentive structure currently discourages the behavior.

How Employers Can Address Performative AI Without Making It Worse

The reporting this week doesn’t offer a prescribed fix from Visier itself, but the pattern across all the coverage points toward a few sensible, low-risk moves for employers who want to close the gap between claimed and real AI usage rather than widen it. Reducing blanket “use AI more” messaging in favor of specific, task-based guidance removes some of the ambiguity that drives exaggeration in the first place. Normalizing admissions of “I haven’t figured this tool out yet” in team settings, rather than treating it as a performance red flag, removes the fear that’s clearly driving a chunk of the 45% pressure figure.

Just as important: tying AI-related performance expectations to concrete outcomes rather than tool-usage claims gives employees nothing to gain by exaggerating. If the metric that matters is “did the report ship faster and with fewer errors” rather than “did you mention using an AI tool,” the incentive to perform fluency largely disappears on its own.

Frequently Asked Questions

What is “performative AI” at work?
It’s a term used by workforce analytics firm Visier, reported by TechSpot, to describe employees who exaggerate or fake their use of AI tools, or their expertise with them, in order to appear more capable or more in step with workplace expectations to colleagues and managers.

How many employees admit to faking AI use at work?
According to the Visier data reported by TechRadar on October 12, 2026, 48% of employees said they exaggerated their AI usage or expertise to colleagues or managers “some or most of the time.” An earlier, separate 2025 survey from Howdy.com, reported by The Register, found about 16% admitted to outright pretending to use AI to please their bosses.

Why do employees feel pressured to use AI?
The Visier survey found 45% of employees feel pressured to use AI at work. Related reporting from The Register notes that roughly three-quarters of employers expect some form of AI use from staff, with about half of those expectations formalized and the rest left informal.

How large was the Visier survey?
TechSpot reported the survey covered 1,000 full-time US employees.

Is performative AI the same as general workplace dishonesty?
Not exactly. It specifically refers to exaggerating familiarity or usage of AI tools, rather than general misrepresentation of hours, effort, or output. Reports frame it as a distinct AI-era behavior tied to fear of appearing behind on a specific, highly visible technology trend.

Are employees worried about losing their jobs to AI?
Separate 2026 survey data from Howworks.ai, cited by 36Kr, found 51% of American workers are worried about losing their jobs to AI, and 72% say they lack confidence in their ability to adapt to AI tools.

What can employers do about performative AI adoption?
Reporting on the trend suggests shifting away from self-reported usage surveys toward outcome-based metrics, reducing ambiguous “use AI more” mandates, and normalizing honest admissions of unfamiliarity with specific tools, so employees have less incentive to exaggerate their fluency.

Does this affect how companies measure AI ROI?
Yes. If a meaningful share of self-reported “active users” are inflating their actual usage, internal adoption dashboards and license-utilization reports used to justify AI software spend become less reliable, pushing more IT and HR leaders toward task-level verification instead of survey-based reporting.

Related Coverage

• Nvidia's China Share Falls to 34%, Huawei Hits 28% [2026]
• Anthropic Cuts AI Distillation Attacks by 99% [2026]
• Meta Stock Jumps 4% on Muse Wonder Frontier Bet [2026]
• Nadella's 137B-Param AI Makes PC a Software Factory [2026]
• Credo Launches SRMO Optics Standard With 7 Partners [2026]