# Story 5: Duolingo's AI-first revolution

- **Episode:** [226](https://agents.truthliesandwork.com/episodes/226.md) (2025-08-26)
- **Type:** other
- **Tags:** ai, duolingo, innovation
- **Time:** 37:46–46:03 (start 2266s)
- **Listen:** [https://truthliesandwork.com/226?t=2266](https://truthliesandwork.com/226?t=2266)

---

[37:46] Leanne Elliott: I will. Number 3, Duolingo's AI-first revolution
[37:52] Al Elliott: Oh, I don't know this, I don't think
[37:53] Leanne Elliott: Oh, you do
[37:53] Al Elliott: Do I?
[37:54] Leanne Elliott: I'm sure you do. So, this has kind of been happening all year. So, Duolingo, you know, the app like you learn different languages on, their CEO, Luis von Ahn, announced that the company would gradually stop using contractors to do work that AI can handle and become AI-first. So, one of the first big organisations to really come out and say, make this claim, they're going to be AI-first. So, Duolingo has already replaced 100 workers, primarily writers and translators, Who create all the kind of quizzes and learning materials that the company is very well known for. This actually followed earlier cuts in January 2024, when Duolingo cut 10% of its contractor workforce, again, as the company shifted content creation to AI. So, the CEO, von Ahn, said teams must prove they can't automate work before hiring, and candidates' AI facility will be evaluated during hiring
[38:51] Al Elliott: Hmm
[38:52] Leanne Elliott: So, basically what they're saying is that if I'm a hiring manager and I want to make a hire, prove that AI can't do it before I give you that budget. A former worker said that the AI output is very boring. Duolingo has always been known as a very fun brand, very quirky, about, yeah, just kind of being that lovely app, isn't it, that helps people learn and engage. But it has received criticism for some mistakes coming out currently
[39:20] Al Elliott: Yeah
[39:20] Leanne Elliott: Or recently
[39:21] Al Elliott: In the content
[39:22] Leanne Elliott: In the content, yeah. Which, after this criticism, again, the CEO walked back on his comments and said that, I do not see AI as replacing what our employees do, but potentially augmenting it. So, yeah, there was a journalist, Brian Merchant, who pointed to this as part of an AI jobs crisis. That's a series of management decisions being made by executives seeking to cut labor costs. And other users have criticised the company online, with many threatening to delete the app over quality concerns and ethical issues. So yeah, is it an honest innovation, Al? Is it about scaling education globally through the inevitable use of AI, or is it corporate greed destroying the creative job industry? Thoughts?
[40:05] Al Elliott: I don't think that's very particularly creative to translate something. Oh, I think, I think you just translate something. I think if they were writing stories, Then that's creative. But no. So no, I'm not gonna, I'm not gonna class that as creative. Do you know what? There's, have you seen this?
[40:20] Leanne Elliott: What level have you got up to on Duolingo though? 'Cause I'll be honest, I've not really got past the kind of general quizzes. Does it get a bit more complex as it goes on?
[40:27] Al Elliott: Buenos días. ¿Dónde está mi coche? That's about as far as I've got. But do you know what? What's funny about this? When I saw this, I do remember this story now. And there's a meme that you see quite often, like these sort of like cheeseburger or whatever you see these websites and it says, me going to bed at 2 AM. And then it's just got, it's got the Duolingo owl or whatever it's called thing going, let's review your mistakes. And I like the idea that whether this CEO, that he's basically someone's done memes about him going, let's review your mistakes without outsourcing to AI. Look, I'm going to be really—
[41:04] Leanne Elliott: Not surprising
[41:05] Al Elliott: I don't know what I've been drinking today, but I'm really grumpy today. AI's here. Just get on with it. So AI took their jobs. So what? Boo-hoo. That's very sad for the individual people. But from an actual, like, generally— Leanne's looking at me going, come on, Al, pull it, rein it in. My point is this
[41:24] Leanne Elliott: I can replace Al. Cancel him, not me, okay?
[41:26] Al Elliott: Basically, you can still say with AI and without, because that's the way my name is. Look, the point is AI's coming for our jobs. AI will take our jobs. If you've ever gone to NotebookLM, I've been listening to podcasts on NotebookLM that they've been creating from documents, and they're pretty good. They're like 60% good, 70% good, better than a lot of other podcasts I've been listening to. We're replaceable. Everyone's replaceable by AI. The point is, that was a shame. But then when, you know, the calculator came in, people who were adding up on a reckoning machine were replaced. It's just the way the progression— don't shoot the man from Duolingo. Be angry. I'm mixing metaphors. Don't blame the game. Don't blame the player, blame the game. Leigh, school me. I'm gonna get some wine from the fridge while you're schooling me
[42:14] Leanne Elliott: No, and you know what? This is, this is a difficult one actually, because I don't, I agree with you. I'm not, I'm not against AI coming into organizations. I think it's happening whether we like it or not. I think it's how it's done. The ethics of it really interests me, and I don't think there's many organizations out there that are really thinking about the ethical implications of this, the societal implications of this, because I think it is, yeah, this is huge. This is like, this is bigger than the internet, I think. This is, this is bigger in terms of its impact on employability. I just think it's a very bold claim at a fairly early stage of AI to make the bold declaration we are going to be AI first
[42:56] Al Elliott: Yeah
[42:57] Leanne Elliott: And I think from a marketing and branding perspective, that was the mistake because I think putting it so publicly, it's not necessarily about being innovative because you're not using a whole new AI thing that doesn't exist anymore. You're using an LLM probably, or something similar to translate stuff. So I don't think it's the flex that the CEO thought it would be
[43:19] Al Elliott: Oh no, I might have changed my mind slightly because I didn't know it was a flex
[43:24] Leanne Elliott: Well, I dunno. It's like, why else would you come out and say it?
[43:26] Al Elliott: Oh, you're changing something
[43:29] Leanne Elliott: Because if you're laying people off for it
[43:31] Al Elliott: Ah, now, yes. If he's come out and said, sorry, I've totally interrupted you there, but I won't
[43:36] Leanne Elliott: Carry on. No, no, please, please
[43:37] Al Elliott: My point was that if he comes out and goes, yes, we're going AI first, and as a result, we've got rid of all these people and look, it's all AI. Aren't we clever?
[43:44] Leanne Elliott: Yeah
[43:45] Al Elliott: No, you're a knob
[43:46] Leanne Elliott: Well, that's what the story was then
[43:48] Al Elliott: In that case then, I'll either edit what I've just said out or I just won't bother. But yeah, I've changed my mind. And here we go. Do you know Jamie Dimon? They are going to use— people are going to use what I said about 5 minutes ago as my rant. This will be my Jamie Dimon moment of going, oh, well, boo-hoo, just get on with it. There you go. Kidding, I'm going to context. Anyway, Leanne, would you like another glass of wine?
[44:11] Leanne Elliott: Yes, please. Yes, please. Yeah, I do know. I think we're both on the fence a bit on this one. I think it's how, again, it was how it was handled. The flex just didn't land right. The ethics of this are barely being figured out. They're barely being discussed, let alone figured out
[44:27] Al Elliott: True
[44:28] Leanne Elliott: Then we have kind of the element as well of, of kind of, this is fairly, a fairly new shift in transition. Legislation could potentially come out. I just think to be so bold about it when you're not an AI company. I don't think that's necessary, and I don't think it's gonna make people feel better for losing their jobs because that should— we're just going, it's not personal, I'm going AI first. Oh, well, that's all right then
[44:52] Al Elliott: You know, okay, I have done a U-turn. I have done a U-turn, and I'm gonna say this is an F. Oh, I'm gonna say it's an F because it's— this person is flexing. That's not cool. I think that— I don't think— so what I've changed my mind about is the story is about the, the CEO going, Yeah, but we're going AI first. Isn't that cool? Whereas if the story was about the CEO replacing humans with AI, it'd be like, well, of course, duh, we're all gonna do that. But if it's about that, then it's an F from me
[45:24] Leanne Elliott: Okay. That was, that was a U-turn. Um, I, I think, I don't think, I'm not sure it's that bad. I think this is happening. I just think how that's been communicated to the world and how it's being managed within the organization is, is, Naive. So I think I'm going to give it a C. I think it's average. I think there's some good aspects because again, it's sparking conversation around people actually replacing people with AI and the problems that come with that, both in terms of quality, quality and ethics and how that brand is perceived within, within the market. Um, yeah, I, I, uh, yeah, it's very, it's average. It was, it was, yeah. Meh