If you've searched anything about Korea this week, there's a good chance you've seen it. "South Korea free AI" has been near the top of Google Trends. I didn't fully understand what it was either, until I looked into it.
 |
| AI-generated illustration of Korea's "AI for Everyone" initiative |
What Exactly Is "AI for Everyone" (모두의 AI)?
The project's real name is "AI for Everyone," run by Korea's Ministry of Science and ICT. The idea is simple: give every Korean citizen a free domestic AI chatbot, with no usage limits. Right now, most people only get that kind of access from foreign tools like ChatGPT or Gemini, and usually with some limits attached.
On August 28, the ministry picked three companies to build it: SK Telecom, Kakao, and KT. These are three of Korea's biggest tech and telecom companies. Each one leads its own group of smaller AI and service companies. The plan moves fast from here. A formal agreement gets signed with all three sometime this September. Then comes a beta test. Then a full public launch, targeted before the end of the year. The government says it wants to keep this free through at least 2028.
There are real rules attached, too. Each service has to use Korean-made AI models for at least half of what powers the chatbot. So this isn't just a Korean-language wrapper on a foreign model — it has to actually run on domestic technology. And it's not only a chatbot. The government also wants a "public AI agent" built in, one that can find government services a person qualifies for and help them apply, instead of just answering questions.
• Selected companies: SK Telecom, Kakao, KT (announced Aug 28, 2026)
• Beta service: expected this September
• Full public launch: targeted by the end of 2026
• Government support: 512 units of Nvidia's B200 GPUs allocated this year
• Domestic AI model requirement: at least 50% of the core model
When I first heard "모두의 AI," I didn't connect it to ChatGPT or Claude right away. My first thought was more like a local comparison.
If Canva is the well-known international option, this felt more like Miricanvas — the Korean alternative a lot of us actually use, because it's built around how we work here.
That's roughly how "AI for Everyone" landed for me. Not a replacement for the AI tools I already use, but Korea's own attempt to make something free and easy for people who might never have tried AI otherwise.
Why This Is Such a Big Deal Right Now
This isn't a small test project. Korea's AI budget for 2026 is about 10 trillion won, roughly 7.2 billion USD. That's almost three times last year's budget. And the reason isn't just "let's build a fun chatbot." Seoul officials estimate that more than 20 million people in Korea already use free foreign AI tools regularly. That's a huge part of the population depending on services built and priced by companies outside the country. The government wants to lower that dependence, help people who can't or won't pay for AI subscriptions, and make AI a normal part of daily life here, the same way free public wifi is.
There's a catch, though. Tech writers here have pointed it out clearly. Once this launches, people won't use it in a vacuum. It will sit right next to ChatGPT, Gemini, and other paid tools many Koreans already know well. If the free domestic version doesn't feel genuinely good to use, people may try it once out of curiosity, then go back to what they already had. So there's real pressure on these three companies to build something people actually want to open a second time.
Curious about using AI in Korea right now, without waiting for this launch? Most major AI apps — ChatGPT, Gemini, and Claude included — already work here with a normal account. No Korean ID or local phone number needed for the free versions. "모두의 AI" is aimed specifically at Korean citizens as a public service, so its sign-up process will likely look different once it's live. But that's a separate question from what's already available to anyone visiting or living here today.
My Own AI Habits — And Why This News Hit Differently
I use AI every single day. Not sometimes — every day. When I'm writing for this blog, I work with Claude. I use it carefully, though, because there's a limit on how much I can use at once, so I try to be efficient with it. When I'm studying Thai, or when I need to translate something properly into English, I switch to Gemini. And more often than I'd like to admit, I have both open at the same time. I ask them the same question just to compare the answers, or to double-check something one of them told me.
That daily habit is exactly why this news felt personal to me, not just interesting. I already lean on AI heavily, every day, and almost all of it is foreign-made. So when I heard Korea was building its own free version, my first reaction wasn't pure excitement. It was excitement mixed with a specific kind of worry: are we starting too late to catch up?
But here's the other side of how I think about this, and it comes from something I've noticed about Korea in general, not just about AI. We rarely invent the technology first. But once something new shows up, I think we absorb it faster than almost anyone else — and often end up using it better than the country that invented it.
The clearest example, to me, is the internet itself. Korea didn't invent the web. That happened elsewhere. But once it arrived, Korea moved on it faster than almost any other country on earth. By the mid-2000s, Korea already had broadband in the large majority of homes, nationwide, years before countries like Singapore or the UAE built out the infrastructure they're known for today. It wasn't just fast in Seoul, either — it was fast almost everywhere, in a way that even the US still hasn't fully matched outside its major cities. Other countries have caught up on raw speed since then, sure. But Korea got there first, and got there for everyone, not just the capital.
And it's not just about being first. Once Koreans decide they like something, they don't just use it — they go all in, completely, and it's hard for anything else to compete after that. Look at KakaoTalk. It's not the only messaging app in the world, not even close. But in Korea, it holds something like 97% of the market. LINE, which dominates in Japan, Thailand, and Taiwan, barely exists here. Koreans picked KakaoTalk early on, decided they liked it, and that was basically the end of the conversation. I don't think LINE will ever seriously challenge it in Korea, no matter how good it gets elsewhere.
I think AI could follow the same pattern. It didn't start here either. But my honest guess is that, per capita, Korea will end up being one of the countries that actually uses AI the most in daily life, not because we built it first, but because we're fast at picking things up once they exist, and ruthless about dropping what we don't like. I'd bet that even people without much money, as long as they have a phone, end up using some form of AI regularly — probably more than half of the country, eventually. That's just how I see us adopt things.
So yes, I'm nervous about whether "모두의 AI" itself will be good enough. But I'm not nervous about whether Koreans will actually use AI, in general, once it's easy and free to try. That part, I feel pretty confident about.
The Generation Gap: Who Will Actually Use This?
Looking at people around me, I'd say it's split down the middle. Among older adults, the ones who are already curious and open to trying new things pick it up fast. I actively encourage people in that group to use AI — I don't know everything either, and it's genuinely useful for looking things up or getting a second opinion. But for a lot of older Koreans, the real problem isn't the AI itself. It's just getting started. The app, the first few steps, having someone nearby to show them how. If a free, simple, Korean-language chatbot lowers that first barrier, that alone could matter more than how smart the AI actually is.
Younger people are almost the opposite problem, in my opinion. My own child uses AI sometimes, but I actively try to limit it. I worry that if kids depend on AI too early and too often, they stop building the habit of thinking things through by themselves. I don't think banning it completely is the answer. I think it's about order. Let kids build their own thinking first. If they're going to use AI seriously later, or even work in the field one day, I'd rather they understand how it actually works before they start leaning on it to hand them answers.
So when "모두의 AI" actually launches later this year, I don't think the real story will be about the technology itself. I think it'll be about who opens the app a second time. Maybe it's a retired parent who finally feels comfortable asking a chatbot a question instead of asking their kids. Maybe it's someone who tries it once, compares it to what they already use, and quietly goes back. Either way, I'll be watching this one closely. I'll probably write an update once it's actually live and I've had a chance to use it myself.