Performative AI: Why Nearly Half Your Workforce Is Overstating Its AI Skills

Ask almost any leadership team how AI adoption is going and the answer arrives with a number attached: licenses deployed, weekly active users, prompts per seat. A new survey suggests a meaningful share of that number is theater. Visier's research on what it calls "performative AI", based on a survey of 1,000 U.S. full-time employees plus interviews across healthcare, finance, HR, and software, found that 48% of workers have exaggerated their AI usage or expertise to colleagues or leadership at least sometimes. Nearly as many, 45%, say they feel pressure to use AI even when they aren't confident they can use it well.

That is not a story about lazy employees. It is a story about what happens when an organization makes AI use visible and rewarded before it makes AI use understood and supported. People learn to produce the signal leadership is watching for, and the signal stops telling leadership anything true.

The Pressure Is Real, and So Is the Change Behind It

The pressure isn't imaginary. According to Visier's September 23 release, 54% of employees say their role has changed significantly because of AI in the past two years, and 51% say AI has significantly or moderately changed their day-to-day workflow. Seven in 10 are concerned AI could hurt their careers, and 24% name a higher risk of losing their job as their top concern.

Put those together and the exaggeration makes a grim kind of sense. If your role is changing, your career feels exposed, and the people above you are counting AI usage, overstating your fluency is a rational defensive move. One software lead Visier interviewed described the constant question, "why didn't you just have AI do it," as its own source of pressure. Visier's Andrea Derler told CIO Dive that when organizations don't let people show how AI actually fits their work, "they will just give you a blanket answer: 'Oh yeah, I'm using AI for everything, and everything is better.'"

Jacob Appleton, Visier's director of information security, framed the risk as a gap in pace. "If an organization's rushing ahead of where their employees are, or pushing their employees to rush ahead of where leadership is, then you have these risks building up," he told CIO Dive. The risk is not just bad data on a dashboard. It is employees quietly using tools they don't fully understand on work that matters, while claiming more command of them than they have.

Leaders Can't See It, and Training Isn't Filling the Gap

The most uncomfortable numbers in Visier's data are about visibility. Only 28% of employees believe their leaders understand how they use AI today. Only 33% think leaders understand what training and support they need, and 29% think leaders recognize their concerns about AI. On strategy, 32% say their employer lacks a clear AI plan and another 27% aren't sure one exists, so 59% of the workforce either doesn't believe in the plan or can't find it.

Training isn't compensating. Just 21% strongly agree their employer provides adequate AI training, resources, and support. CIO Dive's reporting on the study adds that 41% of workers trained themselves on AI tools, while only 15% credited employer-sponsored training. That combination is the breeding ground for performance: people are self-taught, unevenly skilled, under watch, and unsure what the organization is actually trying to achieve. Asked what would help, employees pointed first to more AI training and upskilling (32%) and greater transparency from leadership (31%), followed by career advancement opportunities (24%) and clearer communication about AI's impact (18%).

"AI adoption cannot simply be measured by tool usage," Derler said in the release. That line should land hardest in HR, because HR is often the function assembling the adoption metrics that go to the executive team and the board.

Managers Are the Hinge, and HR Doubts Them

If there is a single lever that separates real adoption from the performed kind, Gallup's research on AI and workplace culture points to the manager. In its Q1 2026 study of 23,717 employed U.S. adults, employees who strongly agreed their manager champions AI were far more likely to report transformational changes to their work, 33% versus 4% for everyone else. Among employees whose managers actively support AI use, 31% said their culture had improved, compared with 21% among those without that support.

Yet in the same research, Gallup's survey of 102 CHROs at global Fortune 500 companies found half are not confident in their managers' ability to guide employees' AI use. Only 57% provide AI training specifically for people managers, and 62% are building centers of excellence or internal AI champions to carry the transition. The person with the most influence over whether AI use is honest or performed is the person HR trusts least to handle it.

That points to a practical reframe. Performative AI is a change management failure more than a technology one. Employees were told what to use before they were told why, shown how, or given a safe place to say they were struggling. Fixing it means treating AI rollout the way HR would treat any major organizational change: a clear case for the change, managers equipped to lead it, honest two-way feedback, and metrics that measure outcomes rather than compliance.

• Stop reporting usage as adoption. Log-ins and prompt counts reward the performance Visier describes. Pair any usage metric with outcome measures, such as cycle time, error rates, or quality reviews, for the specific workflows AI is supposed to change.
• Publish the plan, even if it is incomplete. With 59% of employees unable to confirm a clear AI plan exists, a plain statement of what the organization is trying to achieve, what is changing, and what isn't will reduce the defensive exaggeration that fills the vacuum.
• Make it safe to say "I don't know how." If admitting low confidence feels like a career risk, people will hide it. Build structured, low-stakes channels for employees to flag where AI isn't working for them, and make sure managers respond without penalty.
• Train managers first, and specifically. Gallup's data shows manager support drives real change, yet half of CHROs doubt their managers can guide AI use. Give managers concrete guidance on evaluating AI-assisted work and coaching people through role changes.
• Close the training gap with the workforce's own priorities. Employees asked first for more AI training and more leadership transparency. With only 21% strongly agreeing current support is adequate, fund role-specific training rather than generic tool tutorials.
• Connect AI capability to career paths. Seventy percent of employees worry AI will hurt their careers, and 24% want clearer advancement opportunities. Showing how genuine AI skill translates into growth removes much of the incentive to fake it.

The organizations that get honest AI adoption will not be the ones with the highest usage numbers. They will be the ones where an employee can tell their manager the tool isn't helping yet, and hear "show me where" instead of "use it more."