The Scoop: Duolingo CEO walks back ‘AI-first’ memo

Duolingo CEO Luis von Ahn turned to LinkedIn to issue a clarifying statement/pseudo apology after facing backlash over an AI strategy memo that left employees and the public concerned and even angry.

In the original all-hands email from late April, von Ahn announced the language app would become an “AI-first” company and would “gradually stop using contractors to do work that AI can handle.” He added that employee headcount would only increase “if a team cannot automate more of their work,” and the company would begin applying AI in hiring and performance reviews.

The tone of the memo was seen as abrupt and dismissive of people’s roles, sparking criticism, especially from longtime contractors.

In his response, von Ahn admitted: “When I released my AI memo a few weeks ago, I didn’t do that well.” His follow-up looked to add nuance: “I do not see AI as replacing what our employees do (we are in fact continuing to hire at the same speed as before). I see it as a tool to accelerate what we do.”

Acknowledging the uncertainty AI creates, he framed the change as proactive and opportunity-driven, not a workforce reduction. To support teams, Duolingo will offer workshops, advisory councils and time for experimentation, he said, noting that the company wants to help employees feel “empowered and prepared,” while reaffirming that people – not just tools – remain central to the company’s mission.

Why it matters: The Duolingo memo saga shows how quickly a message can spiral out of control – and how critical it is to get tone right from the start.

The original AI memo wasn’t leaked; Duolingo posted it publicly on LinkedIn, which is smart in theory since “internal-only” no longer really exists. But transparency alone isn’t enough. The memo’s clinical tone and lack of empathy for the impact on contractors and employees left a void filled by anxiety and negative headlines.

As one LinkedIn commenter put it on Von Ahn’s latest post, the first memo did “untold damage” to the company’s reputation, and “no amount of walking its inflecting AI statement back will undo the perception that people have of it.”

Von Ahn and his leadership recognized they needed to respond – not just for optics but to rebuild trust. His follow-up acknowledged the fallout and emphasized that no one is expected to navigate this shift alone.

Von Ahn made a smart play to highlight the importance of frontline staff, saying the people work at Duolingo because they want to “solve big problems to improve education,” and that the people who work there “are what make Duolingo successful.”

The major takeaway here is that while transparency is important, it needs to come in tandem with context. And in this case, that was considering the human impacts of going “AI-first.” In times of big change, such as anything involving jobs, make sure messaging is clear and human-focused, over-communicate if you have to. Never assume your audience will fill in the blanks the way you want them to.

Editor’s Top Reads

• At the Cannes Film Festival, 34 TikTok creators with filmmaking ties gathered at a festival hub, where Tom Cruise surprised them with a fireside chat and one-on-one time. Creators like Reece Feldman, known for his behind-the-scenes videos, and Zainab Jiwa, who went viral for her Denzel Washington interview and hosted red carpet moments, shared their personal festival experiences. A clip of Feldman’s interview with Cruise racked up 5,000 likes in the first 17 hours. “What I’ve tried to do in every step of my journey is to take the audience with me, because I never had that growing up,” Jiwa told Variety. For PR pros, when you’re working with creators who have authentic, engaged communities – not just big follower counts – and give them meaningful access. People are following them for their content, not just to see movie stars. As Feldman says, social media should be “approachable” and a platform for voices beyond the usual stories.
• Phishing scams posing as state DMVs are spreading, sending urgent texts demanding payment for fake tolls or violations. These messages often abuse real legal codes and create a false sense of urgency. To combat this, states are using a mixture of earned media and official government channels ensure people have a mix of platforms to turn to for reliable information. “The best way to protect yourself is to always use the DMV website or call our contact center to interact with us,” California DMV Director Steve Gordon wrote in a statement. His warning was sent to newsrooms across the state to add credibility. In New York, the DMV issued a social media alert: “Bad actors are sending texts that claim to be from DMV telling people they must immediately pay fines. Please don’t give personal information to these scammers,” with a link to its official site to help people verify the message. The best way to fight scammers and false info? Be louder, faster, and clearer than they are. Get your message on as many trusted channels as possible and make it easy for people to spot what’s real.
• A Swedish startup called Fibbl is giving product pages a major glow-up by replacing static photos with interactive 3D models that customers can spin, zoom and explore using AR. Brands like Samsonite reported early tests showing that customers who engaged with these models were 50% more likely to convert. “Interaction is life,” Edouard Wattel, Samsonite’s VP of digital and e-commerce, told NBC News. For PR and marketing teams, this isn’t just a cool tech upgrade, it’s a genuine storytelling opportunity. AR and VR have long been hyped but often stayed stuck in flashy demos. But this new data shows that brands can create practical, scalable and usable across digital platforms – from websites to video games. It lets consumers virtually touch and interact with products, making brand engagement more memorable. It’s a reminder to stay open to new ways to offer potential customers fresh, (almost-)real ways to engage with your content.

Casey Weldon is a reporter for PR Daily. Follow him on LinkedIn.