Duolingo CEO and cofounder Luis von Ahn just made a statement on AI tools at his company, saying that he doesn’t “see AI as replacing what our employees do.”
This view certainly appears to be a reversal of the company’s stance one week earlier, when the CEO said that the company was going to “stop using contractors” due to a major shift that would make Duolingo an “AI-first” company.
This about-face isn’t unfamiliar: Companies including Klarna, Starbucks, and others have walked back similarly aggressive pivots to AI. What’s behind this trend? Public backlash, failures to deliver, or something else entirely?
Why Duolingo Says It’s Still Hiring at the Same Pace
For weeks, Duolingo has been making noise about the benefits of AI. That’s no surprise to anyone following its 2025 hiring practices: The educational technology company laid off 10% of its workforce in January, saying it was phasing out human-led translation, to replace it with AI tools.
Most recently, however, Duolingo made its most AI-positive move yet, saying a week ago that it would “gradually stop using contractors to do work AI can handle” and it was committing to being “AI-first” as a company.
This just in! View
the top business tech deals for 2025 👨💻
Now, though, the company is walking back these comments. The company CEO explained in a LinkedIn statement.
“To be clear: 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, at the same or better level of quality.” – Duolingo CEO Luis von Ahn
The CEO also highlighted the “rigorous standards” that all AI-generated content would be held to.
It all seems to be a reversal, given that the company said it was going to “stop using contractors” and is now saying that it is “continuing to hire at the same speed.” It’s possible that the hiring refers to payrolled rather than contract workers. However, Duolingo is far from the only business to reverse course on AI replacement.
Why Klarna Is Backtracking
The Swedish buy-now-pay-later fintech company Klarna has also waffled on AI tools in a public manner. CEO Sebastian Siemiatkowski is a huge AI proponent: Just this week, he used an AI avatar of himself to deliver part of the company earnings call.
Earlier in the month, Siemiatkowski told CNBC that AI had helped the company “streamline its workforce by ~40%” and that the company had shrunk from “about 5,000 to now almost 3,000 employees.” Those cuts included 700 customer service workers… which the company now regrets cutting.
Recent statements from Siemiatkowski include:
- “I just think it’s so critical that you are clear to your customer that there will be always a human if you want.”
- “As cost unfortunately seems to have been a too predominant evaluation factor when organizing this, what you end up having is lower quality.”
- “Really investing in the quality of the human support is the way of the future for us.”
Why turn from promoting AI to highlighting the value of human contributions? Possibly to boost valuations before Klarna IPOs. The company was worth $45.6 billion in 2021, which has now dropped to $6.7 billion today. The company’s dramatic workforce reductions began in 2022, so investors may have concerns that AI hasn’t been able to maintain value as well as Siemiatkowski expected.
Why Starbucks Stopped Replacing Workers With AI
In an investor call earlier this month, the Starbucks CEO put it plainly.
“What we’re discovering is the equipment doesn’t solve the customer experience that we need to provide, but rather staffing the stores and deploying with this technology behind it does.” – Starbucks CEO Brian Niccol
The company had been “removing labor from stores” over the last few years, assuming that the tech could fill the lack of workers. Now, Niccol says, “what we’re finding is that wasn’t an accurate assumption with what played out.”
It’s about as straightforward as any CEO is likely to get: Across the last two or three years, many big bets on AI haven’t paid off as well as they would have liked.
What’s Behind All the AI Pivot Reversals?
It may not be possible to break down a massive, multi-industry shift towards and back away from AI. If you were to try, however, these are likely a few of the top reasons that you might come up with.
All for PR?
One reason might be that a given company’s entire announcement and reversal was always more of a PR move than a business one.
Duolingo is a social media marketing heavyweight. A recent Link in Bio survey of social media professionals ranked it number one for “great social,” and the brand is constantly earning headlines about odd stuff like how its owl mascot has just come back from the dead. And Duolingo definitely earned plenty of social media attention with the initial announcement, posted on the suprisingly buzzy LinkedIn platform, and they seem set to get more attention with their change of mind.
Backlash proved too much?
Part of rapid adaptation is knowing when to pivot again, and customer backlash is a big reason to do just that. Plenty of customers are anti-AI for a host of reasons. Some hate the copyright violations that companies like OpenAI say are essentially inevitable, while others hate the errors that chatbots constantly churn out or note the increasing energy costs are harming the environment.
But the main reason is the impact on the labor market: Inflation and the nearing recession are already squeezing workers, so the threat of AI taking their jobs is particularly stinging.
Learning pains?
The most measured reason why a company might make a big change and then walk it back: They bit off more than they could chew. Dramatic shifts in direction are tough, since they require unfamilar practices. The backtracking could be a natural result of the new technology.
Klarna’s failures to keep up with customer service requirements emerged due to their AI reliance, so the company has learned a lesson and will likely move forward with a better balance of AI to human workers. Starbucks seems to be saying the same.
AI Adaption Remains a Huge Movement
Ultimately, the drive behind AI adaption is still huge: Saving money and surpressing the labor market are both great ways to grow a company.
Klarna, Starbucks, and Duolingo are admitting that AI can’t do everything, but their actions over the last few years indicate that they’re still interested in getting AI to do as much as possible. That’s unlikely to change, even if they are constantly admitting they’ve gone too far.
These companies, along with the rest of the corporate world, are committed to the efficency benefits of AI. They still have a lot to learn about the limits of those benefits, however, and they’ll be slowly learning more and more about what AI still can’t do for years to come.
