For three years, every company that deployed AI tools said the same thing. Adoption is encouraged. It is a personal choice. Use it if it helps. Skip it if it does not. The implicit promise was that AI was a resource, not a requirement. Your performance was still measured by what you delivered, not by which tools you used to deliver it.
Meta ended that premise on February 4, 2026.
The company’s Head of People rolled out a new performance review framework. “AI-driven impact” became a formal metric. Not a suggestion. Not a bonus category. A graded expectation, evaluated in the same cycle that determines promotions, compensation, and whether you still have a job next quarter.
Then came the detail that makes the whole thing recursive. Meta told employees to use AI to write their performance self-reviews. The tool you are being graded on adopting is the tool you are expected to use to write the document proving you adopted it. Compliance generates the evidence of compliance.
The timing is the part no one can look away from. In the same period Meta codified AI usage into every employee’s review, the company cut roughly 8,000 positions. The performance framework did not arrive in a vacuum. It arrived alongside a workforce restructuring that made clear what happens to people who score poorly on the new metric.
This is not a productivity initiative. It is a sorting mechanism. The review does not ask whether AI made your work better. It asks whether you used it. The distinction matters. One measures output. The other measures obedience. And obedience is measurable at scale, which means it can be automated, which means the grading system itself is a filter designed to identify who the next round of cuts should remove.
Every other company is watching. When the largest social media company on earth makes AI adoption a condition of employment, it gives permission to every mid-market CEO, every VP of Engineering, every Head of People at a company one tenth the size to do the same thing. The template is set. The question is no longer whether your company will grade you on AI usage. The question is whether they have told you yet.
Meta’s Head of People framed it as empowerment. The review cycle framed it as a requirement. The 8,000 cuts framed it as a consequence.
The optional era lasted three years. It ended in a performance review.
👉 Read the agentic security news. Instantly analyze the threat vector, see if it applies to your setup, and find the gaps with our interactive playbook. All free.
👉 Follow me on LinkedIn | X | Substack for weekly analysis of real agent failures, control gaps, and what the frameworks are and are not catching.
Comments
Nothing yet. Say the first thing.
Sign in to join the conversation.