When ChatGPT started writing performance reviews in 2024, nobody was surprised. What surprised people was how good the AI-generated reviews were—often indistinguishable from the originals, sometimes better.

That tells you everything you need to know about the system.

If an algorithm can replicate your annual performance feedback after being fed three bullet points and a job title, the feedback was never about performance. It was theater. Specifically, it was Kabuki theater—elaborate, stylized performance following rigid conventions, where everyone knows their role and the outcome is predetermined.

AI didn’t break performance reviews. It exposed that they were already broken. The system’s rapid adoption of ChatGPT for review-writing reveals an uncomfortable truth: nobody—not managers, not employees, not HR—actually believed in the process. They were just going through the motions.

The $35 Million Charade

A company with 10,000 employees spends approximately $35 million annually conducting performance reviews. Managers invest 210 hours per year. Employees spend 40 hours. That’s one full work week dedicated to documentation theater.

The return on that investment? 30% of performance reviews actually decrease employee performance. Not “fail to improve”—actively harm it.

Research by Betterworks found that 64% of employees consider performance reviews a complete or partial waste of time. Less than one-third view the process as “very fair and equitable.” When employees perceive reviews as biased, they’re 2.5 times more likely to be actively job-hunting.

Apple’s former chief talent officer called annual reviews “the stupidest thing American companies do.” Microsoft, GE, Adobe, and Netflix have all abandoned traditional reviews. These aren’t small startups experimenting—these are Fortune 500 companies concluding that the entire apparatus produces negative value.

Yet most organizations persist. The question is why.

What Performance Reviews Actually Do (Hint: It’s Not About Performance)

Performance reviews exist to serve institutional needs, not employee development:

1. Legal documentation. The primary function is creating a paper trail defensible in wrongful termination lawsuits. Managers aren’t evaluating performance—they’re building a liability shield.

2. Compensation justification. Reviews provide ex-post rationalization for salary and bonus decisions already made during budgeting cycles. The feedback follows the number, not vice versa.

3. Forced ranking. Many organizations require bell-curve distributions. Even if your entire team exceeds expectations, somebody must be rated underperforming because math. This actively undermines collaboration—your success requires your colleague’s failure.

4. Hierarchy reinforcement. The annual review ritual reminds everyone of their place in the org chart. It’s a power ceremony, not a development tool.

5. HR process theater. Someone in People Operations needs to demonstrate they’re adding value. Elaborate review systems with competency matrices and 360-degree feedback look impressive in budget presentations, regardless of whether they improve outcomes.

The stated purpose—improving employee performance through constructive feedback—ranks somewhere around fifth in the system’s actual priorities.

The Information Problem: Why Managers Can’t Actually Evaluate You

Even if reviews were designed honestly, they’d fail due to structural information asymmetry.

Knowledge work is largely invisible. A software engineer’s contribution to system architecture, a product manager’s influence on roadmap prioritization, a data scientist’s insight that prevented a costly model deployment—these don’t show up in dashboards. They’re not captured in metrics. Months later, when review season arrives, they’re forgotten.

Studies show that companies lose 20-30% in revenue annually due to data silos. Knowledge workers waste 12 hours per week “chasing data” trapped across systems. Yet somehow, your manager is supposed to have complete, objective data about cross-functional projects you worked on with four different departments?

Recency bias dominates. Whatever happened in the last 60 days weighs disproportionately in evaluations, regardless of what occurred in months 1-10 of the review period. Managers default to recent interactions because human memory is unreliable and note-taking is inconsistent.

The result: evaluations reflect visibility rather than value. The person who talks loudest in meetings gets better reviews than the person who prevented a production incident at 2 AM. The IC who sends frequent updates to senior leadership outperforms the one solving harder technical problems in obscurity.

Performance reviews measure optics, not outcomes.

The Ratings Fraud

CEB research found that two-thirds of employees receiving the highest scores in typical performance management systems are not actually the organization’s highest performers.

More damning: studies show that more than 50% of a performance rating reflects the traits of the evaluator, not the person being evaluated. Your review score predicts your manager’s rating tendencies better than it predicts your actual contribution.

This creates perverse outcomes:

• Rating inflation. Managers game the system to secure higher compensation for favored reports, making ratings meaningless across teams.
• Predetermined distributions. One company, facing a poor fiscal year, instructed leaders to reduce all review scores by 10% to cut bonus payouts. Performance had nothing to do with it—finance needed to hit a number.
• Subjective interpretation of identical work. Your rating depends more on your manager’s risk tolerance, their political capital, and the stack rank already assigned to your level than on what you actually accomplished.

The system doesn’t measure performance. It measures management theater.

Why Netflix, Apple, and Microsoft Abandoned This

Progressive companies didn’t replace formal reviews with nothing. They replaced them with something that actually works: continuous feedback.

Netflix abolished formal reviews in favor of informal 360-degree conversations year-round. Former Chief Talent Officer Patty McCord explained: “If you talk simply and honestly about performance on a regular basis, you can get good results—probably better ones than a company that grades everyone on a five-point scale.”

Apple eliminated annual reviews because they’re “a tremendous waste of time.” Real-time feedback proved far more effective.

Microsoft dumped stack ranking after realizing it undermined collaboration and innovation. Forcing managers to label 10% of employees as underperformers regardless of actual team performance created a zero-sum culture where helping a colleague could hurt your own rating.

The pattern: companies that compete on innovation and talent retention can’t afford the morale damage performance reviews inflict. Only organizations with monopolistic positions or commoditized workforces can sustain the system’s costs.

If you’re working somewhere that still does traditional annual reviews, that tells you something about how leadership views your leverage.

What AI Reveals: Managers Never Believed Their Own Feedback

When managers started using ChatGPT to draft reviews in 2024, initial reactions focused on ethics: is it appropriate to outsource human judgment to AI?

Wrong question. The right question: why did adoption happen so quickly and with so little resistance?

Because managers knew the reviews were bullshit. They’d been writing them by template for years—”strong contributor… areas for growth… looking forward to continued development.” The specific wording never mattered because the content was predetermined by comp decisions made months earlier.

AI didn’t degrade the quality of feedback. It automated a task that was already mechanical.

The speed with which directors of engineering at companies like BetterUp openly shared their ChatGPT performance review prompts on LinkedIn revealed nobody was embarrassed about AI-generating reviews. Why would they be? The original system was already content-free.

If the process had actual value—if reviews genuinely drove development and growth—managers would resist automation. You don’t outsource strategic thinking to ChatGPT. You outsource bullshit.

The performance review system’s wholesale adoption of AI is an implicit admission from every participant that the emperor has no clothes.

The Real Cost: Opportunity and Trust

The $35 million direct cost understates the damage.

Opportunity cost. Those 210 manager-hours per year could fund actual coaching, mentorship, or 1-on-1s focused on real-time problems. Instead, they’re spent on documentation theater that generates zero insight.

Trust erosion. When employees know reviews are arbitrary, backward-looking, and disconnected from actual contribution, they disengage from development conversations entirely. Why discuss growth with someone who’s reading from a script to justify a predetermined rating?

Risk aversion. If every action might become evidence in a future review, employees stop taking intelligent risks. Innovation requires experimentation. Experimentation involves failure. But failure creates documentation risk. Safer to do mediocre, defensible work than excellent, unconventional work that might not align with next year’s arbitrary competency rubric.

The system doesn’t just waste time—it actively punishes the behaviors that drive outsized value.

What High Performers Should Do Instead

If you’re still subject to traditional performance reviews, treat them as the compliance exercise they are. Don’t mistake the ritual for reality.

Document your own impact. Maintain a “wins file” with specific examples, metrics, and testimonials. When review season arrives, you can feed your manager bullet points that make their ChatGPT prompt easier. This isn’t about helping HR—it’s about ensuring your actual contributions are represented, even badly.

Build visibility through output, not optics. The best defense against subjective ratings is objective results. Ship features customers notice. Solve problems that would otherwise escalate. Generate revenue. Create artifacts (decks, docs, tools) that have your name on them. Make your value legible to people outside your immediate reporting chain.

Optimize for options, not ratings. A “meets expectations” rating at a high-trajectory company beats “exceeds expectations” at a dying one. Your goal isn’t maximizing your current performance score—it’s building skills and relationships that make you unfireable across multiple roles. Focus on work that compounds: technical skills, domain expertise, reputation, network.

Decouple feedback from reviews. Seek real-time input from people whose judgment you trust—peers doing similar work, mentors who’ve navigated your path, customers using your product. Performance reviews happen once a year and are heavily filtered through organizational politics. Genuine feedback happens continuously and tells you what’s actually working.

Understand your manager’s incentives. If your manager is gaming the rating system to get you promoted, great—align with that. If they’re constrained by forced distribution and can only give one person a top rating, understand that calculation exists separately from your performance. Don’t take ratings personally when they were never personal to begin with.

Most importantly: recognize that performance reviews are orthogonal to performance. How you’re rated and how you’re performing are loosely correlated at best, uncorrelated at worst, sometimes inversely related.

The System Won’t Change (So Route Around It)

Organizations will not abandon performance reviews voluntarily. The legal, financial, and political incentives are too entrenched. HR departments have built entire empires around elaborate review processes. Consultants sell seven-figure “performance management transformation” projects. Management theory is deeply invested in the fiction that annual ratings drive results.

AI just gave everyone plausible deniability. Now managers can spend 20 minutes feeding ChatGPT instead of 20 hours writing reviews manually, while maintaining the facade of thoughtful evaluation. The system continues with less friction, not more scrutiny.

But you’re not powerless. You can route around broken systems by:

1. Choosing employers who’ve abandoned reviews. Companies that recognize the process is theater signal something about their broader relationship with employees. They either trust you to perform without elaborate monitoring, or they give real-time feedback when it matters. Both are better than annual kabuki.

2. Building leverage. The antidote to arbitrary ratings is making yourself indispensable in ways the organization can’t ignore. Work on critical path projects. Own systems nobody else understands. Develop relationships with decision-makers who can advocate for you when ratings are debated. Become unfireable not through perfect reviews, but through irreplaceable value.

3. Treating comp as market-determined. Your compensation is ultimately determined by what other companies will pay you, not your internal performance rating. Regularly test the external market. Interview annually even if you’re happy—it gives you data on your actual value vs. what your current employer’s rating system claims.

4. Focusing your energy where returns are non-linear. Performance reviews produce linear outcomes—0.5% more raise, 1 level earlier promotion. High-leverage work produces exponential returns: the key insight that redirects a project, the relationship that opens a new domain, the skill that compounds across multiple roles. Don’t optimize for the review rubric. Optimize for capabilities that create asymmetric upside.

The Pattern Recognition

Here’s what AI’s performance review adoption reveals about corporate America more broadly:

When institutions rapidly adopt automation for a supposedly human-centered process without resistance, it’s because participants never believed the process mattered. They were performing for observers, not executing strategy.

This applies beyond performance reviews. How many other corporate rituals survive because they provide legal cover, reinforce hierarchy, or justify existing spend—not because they achieve their stated objective?

Strategic planning cycles that produce binders nobody reads. All-hands meetings that convey no information. Mandatory training that changes no behavior. Town halls where questions are prescreened and answers are vetted by legal.

Performance reviews are just the most visible example of organizational theater masquerading as management practice.

The Bottom Line

Performance reviews are broken by design, not accident. They’re not designed to improve performance—they’re designed to document decisions, manage legal risk, and maintain control. The process survives because it serves institutional interests, not employee development.

AI didn’t expose this by competing with human judgment. It exposed it by easily replicating what was already mechanical. If your annual review can be generated by feeding ChatGPT three bullet points, the review was never about your work. It was about compliance theater.

The question isn’t whether AI should write performance reviews. The question is why we’re still pretending the exercise has value for anyone except Legal, Finance, and HR.

High performers understand this and route around it. They build legible value, maintain external market leverage, and don’t confuse ratings with reality. They recognize performance reviews for what they are: an elaborate charade that consumes resources while pretending to generate insight.

The best response isn’t reforming the system. It’s treating it as the compliance checkbox it is and investing your energy where returns are actually non-linear.