April 29, 2025 Duolingo, the language learning company, is moving to an AI-first operational model, replacing hundreds of contract workers with artificial intelligence. The company said the decision reflects a mature use of AI technology, positioning independent AI agents as viable employees for certain content tasks.

Founded in 2011, Duolingo offers gamified language learning in over 40 languages, serving more than 88 million monthly active users. The company makes most of its revenue through subscriptions, advertising, and English proficiency tests. It has integrated artificial intelligence into its platform for years, including launching “Duolingo Max” in 2023, powered by OpenAI’s GPT-4.

Duolingo CEO Luis von Ahn explained the move in a recent interview with The Verge, stating that the company had bet big on mobile in 2011 and embracing that wave of technology had been a big part of their success.

Today, he feels that AI offers another way to leverage a new wave of technology.

The layoffs primarily affect contractors responsible for content creation tasks such as writing example sentences and reviewing translations. However, Duolingo emphasized that human reviewers are still essential, especially for new courses and lesser-known languages where AI performance can lag.

While many companies have announced AI-first strategies prematurely, Duolingo’s shift is seen as a grounded example of AI replacing repetitive and scalable human tasks. von Ahn also emphasizes that this doesn’t mean that the company is any less dedicated to its employees.

In his email to staff, which was also posted to Linked In, von Ahn says “this isn’t about replacing Duos with AI.” Duolingo will “gradually stop using contractors to do work that AI can handle,” and that the company will become “AI-first.”

This approach contrasts with earlier, more speculative uses of AI, where promises outpaced the real-world capabilities. But AI models have advanced significantly this approach shows an increasing maturity in tackling even a bold approach.

But it is a key step to companies using artificial intelligence not just for experimental features, but for core operational functions, even if in this case the company is reducing AI contractors if not AI employees. As AI systems become more capable, organizations like Duolingo are showing how blended human-AI workflows can be structured to maximize efficiency without fully eliminating human input.