Duolingo Listens to Workers, Drops AI Performance Metrics

In a workplace win for common sense over buzzwords, Duolingo reversed a controversial plan to evaluate employees based on how much they use artificial intelligence.

CEO Luis von Ahn introduced the AI measurement idea as part of performance reviews, but workers quickly pushed back. They asked a simple question: were they supposed to use AI just for the sake of using AI?

Von Ahn listened. "At the end, we backtracked, and we said, 'No, look. The most important thing in your performance is you are doing whatever your job is as well as possible,'" he explained on the Silicon Valley Girl podcast.

The shift reflects a maturing understanding of AI in the workplace. While AI helped Duolingo launch 148 new language courses, the CEO acknowledged that forcing the technology where it doesn't fit actually hurts productivity.

Von Ahn pointed to real examples. AI struggles to create multiple narratives, often producing nonsense. It can't code better than skilled programmers yet, and when AI-generated code has bugs, finding and fixing them takes longer than writing it by hand in the first place.

"It felt like, rather than being held accountable for the actual outcome, we were trying to push something that in some cases did not fit," von Ahn said.

Why This Inspires

This story shows leadership done right. Instead of doubling down on a trendy policy, Duolingo's founder chose employee feedback over ego.

The company maintains it has never done layoffs, despite AI integration. Von Ahn sees AI as making individual employees more productive, not as a replacement for human talent.

This reversal came after Duolingo faced backlash last year when von Ahn declared the company "AI-first" and suggested limiting hiring. He quickly walked that back too, clarifying that AI doesn't replace people.

The lesson is spreading beyond one company. Workers are finding their voices matter, and smart leaders are learning that the best tool is the one that actually works.