Workers Are Doing 'Performative AI,' Exaggerating AI Use to Bosses - Business Insider

TEXT START: In the age of AI changing the way people work, some workers feel like they're under a spotlight, proving their skills with the new technology.

The Dissection

The article is not primarily about “performative AI.” It is documenting a labor market ritual in which workers publicly demonstrate allegiance to the technology threatening their economic position. Employees exaggerate usage, managers exaggerate competence, and executives stage adoption for investors. The resulting theater is rational: when AI becomes the declared measure of productivity, refusing to perform enthusiasm becomes evidence of disposability.

The article correctly identifies a trust gap and an organizational measurement failure. But it frames the crisis as a communications, training, and implementation problem. The deeper mechanism is harsher: workers are being compelled to market themselves as compatible with their own replacement.

The Core Fallacy

The text confuses AI-assisted task improvement with worker viability.

Faster completion, more creative solutions, and higher-quality output do not prove that employees become more valuable. Under the Discontinuity Thesis, they can prove the opposite: if AI raises output per worker, firms need fewer workers. The article treats productivity gains as benefits to labor while ignoring the ownership question—who captures the gains and who becomes economically unnecessary.

It also assumes that better transparency, training, and employee influence can stabilize the transition. Those measures may reduce confusion and improve adoption, but they do not defeat P1, P2, or P3. Once cognitive automation is cheaper and superior, no amount of honest discussion preserves a mass labor market whose primary function has been automated.

“Not yet” is doing ideological work here. The absence of total job destruction today is presented as evidence against the jobpocalypse, when it is merely evidence of lag: institutional inertia, managerial experimentation, and uneven deployment.

Hidden Assumptions

• AI competence will remain a durable labor-market advantage rather than becoming a baseline requirement or an automation bridge.
• Productivity gains will be shared with workers instead of being converted into headcount reduction, wage pressure, or increased performance demands.
• Training and upskilling can restore bargaining power after the underlying tasks have been commoditized.
• Leaders can solve the problem through better communication despite having incentives to conceal displacement risk and perform adoption for investors.
• Human employment remains the default destination for improved output.
• Workers can secure themselves by demonstrating AI usage, even though usage without ownership leaves them operating someone else’s capital.
• Surveyed perceptions of benefit are treated as evidence of durable job value rather than temporary assistance during the transition.
• The central failure is organizational trust, not the collapse of productive participation.

Social Function

Classification: transition management, partial truth, and ideological anesthetic.

The article tells a real story: workers are anxious, organizations lack coherent plans, and AI can improve current tasks. But it packages that reality as a manageable workplace-adjustment issue. It encourages workers to become more legible, enthusiastic, and trainable for the system while leaving the ownership structure untouched.

“Performative AI” is therefore not merely workplace theater. It is a compliance mechanism. Workers are taught to signal that they can extract more value from the machine, while the machine’s owners determine whether that additional value justifies eliminating them.

The Verdict

This is an early symptom of productive participation collapse disguised as a management problem. The article observes the smoke but refuses to name the fire: workers are competing to prove they can operate systems whose strategic purpose is to reduce the need for workers. AI training may improve short-term employability. It does not convert a servitor into a Sovereign, and it does not reverse the terminal logic of mass cognitive automation.