Duolingo drops plan to rate employees on AI use
Duolingo CEO says company reversed decision after workers questioned forced AI usage in jobs
Duolingo AI performance review policy has been rolled back after internal concerns over forced adoption of artificial intelligence in workplace evaluations. The change was confirmed by Luis von Ahn during an appearance on the Silicon Valley Girl podcast.
In its initial formulation, the Duolingo AI performance review policy was launched in April 2025 under an AI-first strategy. The objective of this policy was to measure how employees were using AI in their job and use this metric for performance evaluation.
However, the company has decided not to go down this path now. According to Luis von Ahn, employees thought that the policy forced them to use AI when it wasn't necessary.
The Duolingo AI performance review approach raised internal concerns that staff were being pushed to use AI “for AI’s sake.” According to von Ahn, this led to the company reassessing its strategy.
He said, “At the end, we backtracked, and we said no. The most important thing in your performance is that you are doing whatever your job is as well as possible.”
Despite the discontinuation of the Duolingo AI performance review process, the company has not completely closed down AI where it helps increase productivity. It was also part of the previous memo that the company planned to cut down the use of contractors in any job roles that can be performed using AI.
Von Ahn further made it clear that the company continues to hire in the same manner as before and AI is not seen as a replacement for workers.
Duolingo's move to suspend its AI performance review process comes amid other technology organizations that have introduced strict guidelines in the use of AI. According to rumors, firms like Meta Platforms Inc. are implementing usage targets for AI among their workers under the leadership of Chief Executive Officer Mark Zuckerberg.
-
AI clone of Mark Zuckerberg? Meta bets big on ‘personal superintelligence’
-
AI users split into three groups as gap widens
-
Is WhatsApp end-to-end encryption a ‘giant fraud’? Telegram CEO Pavel Durov thinks so
-
South Korea to launch AI smart city pilot projects in Southeast Asia
-
AI may beat bitcoin in decentralisation, study finds
-
UK watchdog rushes to evaluate risks of Anthropic’s latest AI model
-
New ‘GDI’ metric tracks which countries lead AI race
-
Anthropic steals AI spotlight from OpenAI at HumanX
