Why AI Performance Reviews Are a Legal and Retention Risk

Companies from Meta to Accenture are weaving AI use into employee performance reviews. The move is creating legal exposure, employee backlash, and a measurement problem that few managers have solved.

At Gusto, a human-resources platform, managers sort workers into five AI archetypes during quarterly reviews. New users start as "observers." "Integrators" use AI daily and see "better outcomes consistently." The top tier, "amplifiers," invent new use cases and teach coworkers. Scott Helmes, Gusto's chief people officer, said the goal is to standardize how the company talks about AI fluency. "We want to be able to have a coaching conversation," he said.

The grading rubric has already drawn fire. Amazon and Uber tried leaderboards tracking AI use and then backed away. Duolingo scrapped plans to evaluate AI use in performance reviews. Meta faces a lawsuit from laid-off workers who claim the company used an AI system to rank productivity – including how often employees used AI tools – and then cut lower-ranked staff without considering medical or parental leave. Meta has denied the allegations.

General Assembly, a tech-skills training provider, surveyed 500 business leaders in the US and UK. Nearly half had incorporated AI into employee reviews, looking at tool-usage reports, performance improvements attributed to AI, and anecdotes about efficiency gains. Much of that evaluation is subjective, the survey found.

Richard Landers, a professor of industrial-organizational psychology at the University of Minnesota, said companies are offloading responsibility onto workers without clear benchmarks. "There's a lot of drive to push that responsibility onto frontline workers in a way that's not particularly fair," he said. Some companies reward AI use even when it does not improve output, he added. "You can't really reward AI performance because we don't really know what that even is yet, and it also looks very different for different jobs."

Stefan Camilleri, vice president of engineering at Typeform, said he asks engineers: "Now that I've given you a new tool, how much faster are you doing it?" Speed alone is not enough, he said. "I'm expecting more ambitious output from fewer people." Shensi Ding, cofounder and CEO of the AI software company Merge, said she judges employees on the quality of their AI use. An accountant who automated payment tracking and boosted team productivity earned high marks. "If you're teaching other people, then that's a level above what we would expect," Ding said.

Meta told employees late last year it would add "AI driven impact" to performance reviews. Accenture has reportedly started monitoring AI logins when considering workers for top-level promotions. Google told non-technical employees that AI use could appear in their reviews, though managers are not required to evaluate AI proficiency.

A January Deloitte report found that 84% of companies have not redesigned work to match AI capabilities. The focus remains on training and experimentation, not restructuring workflows or career paths. Lori Moffatt, managing director at Waterstone's Leadership and Culture Advisory Services, said the usage conversation is not delivering the impact executives expected. "AI is meant to be transformational. And yet managing it the ways that they're managing today is not creating the transformation that they're looking for," she said.

Marc Cenedella, founder of the career site Ladders, said workers need time to discover meaningful uses for AI. He encourages experimentation, even if early efforts look like "goofing around." "It kind of doesn't matter if it's useless or useful at this point," Cenedella said. "If this was measurable, that would mean it's not an amazing technology."

Accenture (ACN) carries an Alpha Score of 47, a Mixed rating in the Technology sector, reflecting the uncertainty around its AI integration metrics.