Duolingo CEO Ends Mandatory AI Use for Employee Reviews
Following internal concerns about mandating AI adoption across the workforce, Duolingo has confirmed it will roll back plans to factor the technology into its performance review policy.
Having previously signalled the company would be “AI-first” and that employees would be assessed on their AI use in 2025, CEO Luis von Ahn has since said the company no longer intends to uphold the policy.
Speaking on the Silicon Valley Girl podcast, Luis said the company reconsidered its approach after employees questioned the rationale behind the decision.
“At the end, we backtracked, and we said, ‘No. Look, the most important thing in your performance is that you are doing whatever your job is as well as possible,” Luis commented. “A lot of times AI can help you with that. But if it can’t, I’m not going to force you to do that.”
Duolingo’s stance diverges from a growing number of organisations exploring ways to incentivise employee AI use.
A notable example is Meta’s leaderboard of the top 250 AI token users companywide, which allowed workers to monitor how much AI their colleagues were using. The board was removed by employees after two days of use, following the external sharing of its data.
Despite stepping back from tying performance reviews to AI usage, Duolingo continues to deploy the technology where it demonstrably lifts productivity, including through Birdbrain, its machine learning model introduced in 2020.
In the same communication that outlined its now-abandoned plan to mandate AI use, the company also indicated it would reduce reliance on contractors in roles it believes can be replaced by AI.
Luis said the organisation’s hiring philosophy remains unchanged and that AI is viewed as an enabler rather than a substitute for talent. “The reality is it’s not yet the case that AI is better at coding than humans. I think you still really need engineers, and you’re going to need them for a long time,” he noted on the podcast.
He added that AI-written code is difficult to debug and not a reliable method for writing stories for the programme.
A company spokesperson reinforced the point in an interview with Fortune: “Duolingo has used AI for years to personalise learning and expand access. Technology is core to how we build. That includes how we think about AI’s role across our teams. Our teams’ work depends on human judgment, expertise, and creativity. AI tools assist with that work; they don’t make decisions or replace the people building Duolingo. What drives every decision we make is what’s best for learners.”
Allowing for more creative and strategic work
The announcement of Duolingo’s AI-tied review policy drew immediate backlash from employees and users, following a LinkedIn post by Luis outlining a technology-led restructuring of the workforce.
The reaction quickly spilt onto Duolingo’s TikTok and Instagram channels, which were flooded with negative comments and led the company to temporarily remove all content from both platforms, each with millions of followers.
“AI isn’t just a productivity boost – it helps us get closer to our mission,” Luis said in a 2025 company statement, explaining that manually scaling content was no longer feasible and that without automation, producing new educational material could take years.
Even as it explored changes last year, the company maintained that integrating AI was not about replacing employees but about shifting capacity away from repetitive tasks and toward higher-value, creative, and strategic work.
Duolingo’s decision to suspend its AI-linked performance review arrives as many organisations move in the opposite direction, imposing stricter parameters around AI use. Large tech companies like Meta, for example, are considering employee AI usage targets as part of plans to become “AI-native,” according to Business Insider.