Duolingo AI Backlash: What Happened and Why Users Revolted

Short answer: Duolingo's AI backlash began on April 28, 2025, when CEO Luis von Ahn told employees the company would become "AI-first," phasing out contractors for AI-doable work and weighing AI use in hiring and performance reviews. Users reacted by deleting the app on camera, flooding its social accounts with criticism, and unfollowing it by the hundreds of thousands. Duolingo deleted its own TikTok and Instagram history in response, then spent the next year walking parts of the policy back.

This is the first entry in a series Ban the Bots is building on brand AI backlash — cases where a company's public embrace of AI collided with its own customers and users. Duolingo is one of the clearest examples because the reaction was fast, visible, and measurable in follower counts and app store commentary.

What Duolingo actually announced

Luis von Ahn posted a companywide memo to Duolingo's LinkedIn page on April 28, 2025, stating "Duolingo is going to be AI-first." The memo said the company would "gradually stop using contractors to do work that AI can handle," would factor AI fluency into hiring decisions and performance reviews, and would only approve headcount growth when a team could not automate more of its own work.

Von Ahn wrote that "being AI-first means we will need to rethink much of how we work. Making minor tweaks to systems designed for humans won't get us there," according to reporting from Entrepreneur. The memo did not frame this as a cost-cutting move, but it landed at a moment when AI-driven layoffs were already a live public fear, and it offered little acknowledgment of the people whose contract work would disappear.

It is worth separating two events that later got blurred together. Duolingo had already cut about 10% of its contractor workforce in January 2024, mostly translation quality-checkers, as AI models like GPT-4 took over translation-review tasks, per TechCrunch. The April 2025 memo was a broader policy statement about the company's future direction, not a repeat announcement of that same layoff. Critics frequently cited both events together as evidence of a pattern.

Timeline of the backlash

How users and social media reacted

Users treated the memo as a betrayal of what Duolingo was supposed to represent: a playful, community-driven way to learn a human skill from other humans. Coverage from The 74 quoted longtime learners saying the app had "joined the dark side," worried that AI-generated lessons would carry less linguistic nuance and accuracy than human-reviewed content.

On TikTok, some of the platform's biggest engagement came from users publicly deleting the app, often after years of maintained daily streaks — a visible, self-inflicted sacrifice meant to signal how seriously they took the protest. Top comments on Duolingo's own posts included lines like "Mama, may I have real people running the company" and "How about NO ai, keep your employees," according to search-aggregated reporting on the episode.

The reaction crossed into a broader conversation about AI displacing human labor, a topic Ban the Bots tracks closely in our explainer on how AI is changing jobs and in our look at AI's effect on entry-level hiring. Duolingo's contractor cuts, real or perceived, plugged directly into that anxiety.

Why a language app hit a nerve

Language learning felt different to users than, say, automating a customer-service chat. Language is a human, cultural skill, and Duolingo had spent years building trust on the idea that its content was carefully made by real linguists. An "AI-first" pivot read to many users as an admission that quality would take a back seat to speed.

The TikTok blackout and the mask video

Duolingo's TikTok and Instagram accounts, with roughly 6.7 million and 4.1 million followers respectively at the time, were flooded with negative comments within days of the memo. Fast Company and Ad Age both reported that the company wiped every post from both accounts over a weekend in May 2025, going fully dark rather than defending the policy in public.

The brand lost more than 400,000 TikTok followers in the weeks following the memo, per Fast Company's reporting. When Duolingo finally posted again, it did not apologize directly. Instead, it published a degraded, glitch-effect video styled like a hacked broadcast, showing someone in a three-eyed Duo mask and black hoodie complaining about "corporate overlords," captioned "Duolingo was never funny. We were." The tone read as simultaneously defensive, satirical, and evasive — a stunt rather than a statement.

What Duolingo said in response

Von Ahn's clearest public response came in an August 2025 interview reported by the New York Times and TechCrunch. He said the memo "did not give enough context," and that internally the AI-first policy "was not controversial" — the problem, in his framing, was how it read externally.

He also stated directly: "I do not see AI as replacing what our employees do (we are in fact continuing to hire at the same speed as before)," and said Duolingo has "never laid off any full-time employees" as a result of the policy, according to Fortune. That statement addressed full-time staff specifically; it did not undo the reported 2024 contractor reduction or reverse the stated goal of relying on contractors less over time.

No full apology, just clarification

Notably, von Ahn's public comments read as clarifications of intent rather than an apology for the memo's impact on contractors or for how it was received. He continued to defend the underlying strategy, including a company practice of "f-r-AI-days" — Fridays set aside for employees to experiment with AI tools — while softening the framing around jobs and hiring.

One year later: the quiet reversal

Almost exactly a year after the original memo, von Ahn told the Silicon Valley Girl podcast in April 2026 that Duolingo had dropped AI usage as a formal metric in employee performance reviews. He said: "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. 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.'"

He added that the original approach "felt like rather than being held accountable for the actual outcome, we're trying to just push something that in some cases did not fit," according to Fortune's reporting. Employees had reportedly questioned whether they were being pushed to "use AI for AI's sake" rather than because it actually improved their work — a distinction that echoes the same criticism users raised about the product itself.

This was a partial reversal. Duolingo did not renounce the "AI-first" branding, and the shift toward automating contractor-level work was not undone. What changed was one specific, employee-facing mechanic that had drawn internal friction on top of the external backlash.

Part of a bigger pattern: the AI brand backlash

Duolingo is not an isolated case. Buy-now-pay-later company Klarna announced in 2024 that AI had effectively replaced the work of about 700 customer-service agents, with CEO Sebastian Siemiatkowski publicly touting the efficiency gains. By 2025, Klarna was rehiring human agents after customer satisfaction on complex issues dropped, moving to what it called a hybrid, "Uber-style" support model, according to Forbes. Siemiatkowski said the company had "focused too much on efficiency and cost," adding, "the result was lower quality, and that's not sustainable."

Both cases follow a similar arc: a confident public AI-first announcement, a fast and visible backlash from the people the company depends on, and a later, partial walk-back that keeps AI in place while restoring some of what was cut. Neither company reversed course entirely. This is the pattern Ban the Bots is documenting in our ongoing AI backlash coverage — brands that treat AI adoption as a pure efficiency story tend to underestimate how much of their value came from the humans they were automating away.

What this case study teaches other brands

Three lessons stand out from Duolingo's experience. First, the language of an announcement matters as much as its substance — a memo that reads as clinical and profit-focused will be read that way even if the underlying policy is more modest. Second, a brand's most engaged audience is also its fastest and loudest critic; the same TikTok following that made Duolingo a marketing case study became the channel that punished it within days.

Third, walking back an AI policy does not require abandoning AI. Duolingo kept its AI-first direction and its "f-r-AI-days," while dropping the specific performance-review metric that drew internal pushback. Klarna kept AI in its support stack while rehiring humans for the cases AI handled poorly. The common thread: companies that survive an AI backlash tend to be the ones that separate "using AI" from "forcing AI everywhere regardless of fit."

This is the first case study in a series Ban the Bots is building on brand-level AI backlash — see the full running list in our AI brand backlash tracker. Facts above are drawn from named news sources current as of July 2026; some figures, such as exact Instagram follower losses and precise app-store rating changes, were not independently confirmed in available reporting and are omitted rather than estimated.