Professionals exaggerate AI use to impress employers, study finds

A recent study by Visier, a people management software company, found that nearly half of American workers (48%) admit to occasionally exaggerating their familiarity with or use of artificial intelligence in front of colleagues and superiors. This behavior has been dubbed “performative AI.”

The survey, which interviewed 1,000 full-time employees in the United States, was released in September 2026. Its findings shed light on how corporate pressure to adopt AI can distort organizations’ understanding of the technology’s actual use.

What is performative AI?

The term comes from “performance,” meaning a staged act. Performative AI refers to demonstrating AI knowledge or use to meet company expectations, without necessarily applying the technology effectively in daily work.

In practice, this phenomenon can manifest in various ways: an employee mentions AI tools in meetings to seem up-to-date; another accesses a platform solely to meet a usage target; a third claims to use AI for tasks where it has little impact.

Why did this behavior emerge?

Visier attributes the phenomenon to pressure and a lack of clarity. Nearly half of respondents (45%) said they felt pressured to use AI without having the confidence to apply it correctly. Only 21% strongly agreed that their employer provided adequate training and support for adaptation.

Many companies monitor AI adoption through metrics like user numbers or access frequency. Andrea Derler, a researcher at Visier, warns that adoption cannot be measured solely by how many employees use a tool or how often they use it. In such cases, the metric may reflect compliance with formal goals rather than genuine engagement with the technology.

What risks do companies face?

The main risk is data distortion. If metrics show high usage, leaders may assume teams are prepared, even if actual results don’t match this perception. Investment and training decisions may then be based on an incomplete view of reality.

It’s important to note that the study was conducted in the United States and relied on self-reported data from workers. Therefore, the findings cannot be directly applied to the Brazilian market.

How can organizations address this phenomenon?

Participants suggested measures to reduce pressure: AI training and education (32%), transparency from leaders (31%), career growth opportunities (24%), and clearer communication (18%).

According to HR Brew, which covered the study, HR departments can also encourage AI experimentation and create a safe space to discuss mistakes.

Analysts propose an alternative: evaluate completed tasks and decision quality instead of tracking access. For example, measure how much time a team saved by using a tool to summarize reports, with human review of the final output.

Although the concept is new and tied to a specific study, it highlights a key distinction that’s shaping adoption discussions: tracking AI use is different from proving its value in daily work.