Performative AI Use: Why Your AI Adoption Numbers Might Be Fiction
The Gap Between What Leaders Say and What Everyone Else Does
Start with one number. In a recent UK study, 99% of C-suite leaders said they use AI regularly at work. Among non-management employees, that figure drops to 41%. That's not a gap. It's nearly a different reality entirely, reported by two groups sitting inside the same organisation.
Read generously, it could reflect genuine differences in how AI fits each role. Read more skeptically, and it raises an obvious question: if executives are reporting near-universal AI use at a rate this far out of step with the rest of the workforce, how much of that number is real adoption, and how much is leadership performing the same thing it's asking everyone else to perform?
What Performative AI Use Actually Is
Performative AI use is the practice of visibly demonstrating or exaggerating AI usage and expertise to satisfy employer expectations, rather than genuinely integrating the tools into real work. Recent research from Visier puts a number on how common this has become:
48% of employees admit to exaggerating their AI usage or expertise to colleagues or leadership. A separate study found 16% of professionals say they've outright pretended to use AI when they hadn't.
The behaviour isn't really about AI at all. It's a familiar response to a familiar management failure, dressed in new language. When leaders ask for a visible signal of something without clearly defining what the underlying behaviour should look like, people learn to produce the signal.
This has happened with every metric organisations have ever rewarded without clearly defining: lines of code written, hours logged, emails sent. AI adoption has simply become the newest version of a very old incentive problem.
Why It's Happening Right Now, Specifically
The research points to a consistent root cause: employees have been handed a mandate without a map. Nearly half of employees in recent surveys report feeling genuine pressure to use AI, and a similar share say they're unsure whether their organisation even has a clear plan for how AI is supposed to change their job.
Only about a quarter believe their leaders genuinely understand how employees are actually using the tools day to day.
That combination, pressure without clarity, is close to a perfect recipe for performance over substance. An employee who doesn't fully understand what "good" AI use looks like in their specific role, but knows they're expected to demonstrate it, faces a simple choice: admit the uncertainty and risk looking behind, or project confidence and hope nobody asks a follow-up question.
Given how most performance cultures actually reward visible confidence over visible uncertainty, it's not surprising which option a large share of employees are choosing.
The Measurement Trap Hiding Underneath This
There's a deeper problem for HR and leadership here, beyond the honesty issue: most of the metrics organisations use to track AI adoption are easy to game precisely because they measure activity, not outcome. Token counts, login frequency, number of AI-assisted tasks completed. All of these can rise sharply while the actual quality or usefulness of the underlying work stays flat, or gets worse.
This isn't a hypothetical risk. Research into actual AI-driven productivity gains has found real, serious limits that pure usage metrics completely miss. One large-scale survey found that reported productivity gains from AI use rise only up to a point, around three tools used regularly, and then fall off sharply past that, as the cognitive overhead of juggling multiple AI systems starts to outweigh the time saved.
Separate research has found a meaningful share of the time AI saves on a first draft gets eaten up later by rework, correcting AI output that looked polished but didn't actually hold up. An organisation celebrating rising AI usage numbers, without looking at what happens downstream, can end up rewarding exactly the pattern that's quietly costing it the most.
The Connection to Trust, Not Just Technology
This sits closer to a trust and communication problem than a technology one, and it connects directly to a related pattern worth knowing about: a meaningful share of employees also report hiding their AI use entirely, worried that visible competence with the tools will make them look replaceable. Put those two findings side by side and the picture becomes clearer.
Some employees are performing AI use they don't really have, out of pressure to look compliant. Others are concealing AI use they genuinely do have, out of fear it will look threatening to their own job security. Both are responses to the same underlying condition: employees navigating an AI mandate without a clear, trusted explanation from leadership about what it actually means for them.
That's a communication failure before it's anything else, and it tends to compound the same way other forms of unclear or inconsistent leadership messaging do. An organisation that hasn't built genuine trust around what AI adoption is actually for, and what it means for individual roles, shouldn't be surprised when the resulting data, on both adoption and anxiety, turns out to be unreliable in opposite directions at once.
What This Actually Calls For
The fix isn't a stricter AI usage mandate, since that's the exact pressure that produced the performance problem in the first place. It starts with defining, specifically and by role, what effective AI use actually looks like, rather than leaving employees to infer it from a vague expectation to "use AI more."
A sales team and a legal team shouldn't be held to the same undefined standard of AI adoption, and pretending otherwise just pushes both toward whatever signal is easiest to fake.
It also means shifting what gets measured, from activity toward outcome. Fewer dashboards tracking how often AI gets opened, more honest conversation about whether the work it's producing is actually better, faster, or more accurate than before, and whether that holds up once the initial draft gets reviewed properly.
And it means leadership modelling the behaviour it's asking for honestly, including the uncertainty. A leadership team that performs AI fluency as confidently as the 99% figure above suggests, while the rest of the organisation is visibly struggling to keep up, isn't setting a standard. It's setting the exact incentive that produces more performance and less substance further down the organisation.
At AceNgage, employee listening is built to surface exactly this kind of gap: not what your AI adoption dashboard says, but whether employees genuinely feel equipped, and trusted enough to be honest, about how they're actually using the tools you've given them.
Talk to an AceNgage Expert: https://forms.gle/FnMHhFqizwpGEYkF7
FAQs on Performative AI Use at Work
What is performative AI use?
Performative AI use refers to employees overstating how much they use AI tools, or how skilled they are with them, in order to appear compliant with employer expectations rather than reflecting genuine, substantive adoption of the technology into their actual work.
How common is performative AI use?
Recent research found that 48% of employees admit to exaggerating their AI usage or expertise to colleagues or leadership, and a separate study found 16% of professionals say they have outright pretended to use AI they hadn't actually used.
Why do employees exaggerate their AI use?
Research points to a combination of pressure and unclear expectations. Nearly half of employees report feeling pressure to use AI, while a majority are unsure whether their organisation has a clear plan for how AI should affect their specific role, leading many to project confidence rather than admit uncertainty.
Is there a gap between how leaders and employees report using AI?
Yes, a significant one. One recent study found 99% of C-suite leaders report using AI regularly at work, compared to just 41% of non-management employees, a gap wide enough to raise genuine questions about how much of the leadership figure reflects real, substantive use.
How can organisations reduce performative AI use?
Defining specifically what effective AI use looks like for each role, measuring actual outcomes rather than usage activity, and ensuring leadership models honest, realistic AI adoption rather than projecting inflated confidence all tend to reduce the pressure that drives employees toward performing adoption instead of genuinely practicing it.
Comments