When was AI released to the public? The short answer and the real history
ChatGPT went public on 30 November 2022, but AI itself dates to 1956. Here's when each kind of AI reached ordinary people, with dates you can check.
Written byBurhan AqibSenior writer, policy and analysisWhy you can trust The AI Future
Our writers research every product from its official pages, independent tests and user reports, link each claim to its source and correct mistakes openly. Unless an article says otherwise, we haven't tested the product hands-on ourselves. Companies can't pay for coverage or see reviews before they're published.
Read our editorial policy
If you mean the kind of AI you can chat with, the answer is 30 November 2022, when OpenAI opened ChatGPT as a free research preview. That's the date most people have in mind. But AI as a field is much older: the name was coined for a 1956 summer workshop at Dartmouth College, and you've been using quieter forms of AI (spam filters and Siri, for a start) for well over a decade.
So when was AI released to the public? It depends on which AI you mean, and I think that's the part most short answers skip. Here's the longer version, with dates you can check.
Short answer
AI research started in 1956. Consumer AI arrived piece by piece from the 1990s onwards. Generative AI for everyone, the chatbot moment, began on 30 November 2022 with ChatGPT, which, by Similarweb's count reported in The Guardian, had about 100 million unique visitors two months later.
The background: a 70-year-old field with a 4-year-old public face
The word itself is the easiest thing to date. In 1955, John McCarthy wrote a proposal for a summer research project, and the meeting took place at Dartmouth in 1956. Dartmouth's own history page quotes McCarthy saying the workshop would "proceed on the basis of the conjecture that every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it." That's a bold sentence for 1955. It's also roughly the bet the industry is still making.
Nothing from that summer was "released". It was a research agenda, and for decades AI lived in university labs and on very expensive machines. Early programs did reach a few curious outsiders, though. ELIZA, a chatbot Joseph Weizenbaum wrote at MIT in 1966, used simple pattern-matching rules to imitate a therapist, and some people who typed to it reportedly felt it understood them (which tells you more about people than about ELIZA, in my view).
A rough comparison helps here. The internet existed as a research network from 1969, but most people would say it "came out" in the 1990s, when browsers and home connections made it usable. AI followed a similar path, with one difference: its public debut came in a single, very visible week, so it feels newer than it is.
Between the lab years and ChatGPT, AI kept slipping into products without much fanfare. In May 1997, IBM's Deep Blue beat world chess champion Garry Kasparov; going by IBM's account, it could evaluate 200 million chess positions per second.
You couldn't use Deep Blue, of course. You could watch it, and that changed how the public thought about machines.
The first AI that millions of ordinary people talked to was probably Siri. Apple introduced it on 4 October 2011 with the iPhone 4S, calling it "an intelligent assistant that helps you get things done just by asking." Alexa and Google Assistant followed within a few years. If you want a strict reading of "AI released to the public", 2011 has a decent claim.
Real examples: when each kind of AI reached the public
A timeline makes the layers easier to see. The table below sorts the milestones by what an ordinary person could actually do with them.
| Year | Milestone | Could the public use it? |
|---|---|---|
| 1956 | Dartmouth workshop names the field "artificial intelligence" | No, research only |
| 1966 | ELIZA chatbot at MIT | Only people with access to MIT's computers |
| 1997 | IBM Deep Blue beats Kasparov | No, but the match was public |
| 2011 | Siri launches on the iPhone 4S | Yes, for iPhone 4S owners |
| Nov 2022 | ChatGPT research preview | Yes, free for anyone with a browser |
| Mar 2023 | Anthropic's Claude and Google's Bard open up | Yes, Claude via business access first, Bard via a waitlist |
| Feb 2026 | ChatGPT reports about 900 million weekly users | Yes, as an everyday tool |
The ChatGPT launch post is short and plain. "During the research preview, usage of ChatGPT is free," OpenAI wrote, and that one line is probably why it spread so fast: no app to install and no price. According to UBS analysts quoted by The Guardian, "In 20 years following the internet space, we cannot recall a faster ramp in a consumer internet app." For comparison, the same article says TikTok took about nine months to reach 100 million users and Instagram more than two years.
Rivals moved quickly. Anthropic introduced Claude in March 2023 after a closed alpha with partners such as Notion, Quora and DuckDuckGo. Google started opening access to Bard the same month, describing it as "an early experiment that lets you collaborate with generative AI." Bard has since been renamed Gemini.
And the scale now? In February 2026, OpenAI said ChatGPT had around 900 million weekly active users and more than 50 million consumer subscribers, as eWEEK reported. That's a lot of people for a product that's barely four years old.
Benefits and limitations of dating AI to 2022
Picking November 2022 as the release date has real advantages. It's specific, it's verifiable, and it matches the moment AI stopped being a feature hidden inside other apps and became something you open on purpose. For most readers asking this question, I'd say it's the right headline answer.
It has blind spots, though.
The 2022 date hides about 66 years of work, and it makes AI sound like a sudden invention rather than a slow build. That matters when people judge the technology. If you think AI appeared out of nowhere, it's easy to assume it'll keep jumping forward at the same pace; if you know the field went through long stalls (researchers call the funding slumps of the 1970s and late 1980s "AI winters"), you're likely to be a bit more patient with the hype cycle.
The other limitation is that "AI" isn't one thing. Spam filters, the recommendation feed on your phone and a large language model all count, but they arrived at different times and work in different ways. Our explainer on the main types of AI sorts them out, and this generative AI definition covers the chatbot kind specifically.
Practical takeaways for you
Honestly, the practical takeaway is that you've been an AI user for longer than you might think. If you've used Gmail's spam filter, asked Siri for the weather or followed a map app around traffic, you've relied on machine learning for years. ChatGPT just made it visible and conversational.
That has a few concrete consequences. Chatbots are new enough that their habits aren't settled: they still make up facts, and their training data has cut-off dates, so anything time-sensitive needs checking against a live source. The tools also change quickly. A guide written in 2023 about "how ChatGPT works" is likely out of date on features and pricing, even if the basic idea still holds. If you want the plain mechanics, our guide to large language models is a good place to start.
I'd also suggest being a little wary of anyone who claims AI is either brand new or nothing new. Both claims are half true. The ideas are old; the scale and the free public access are recent.
The next few years
This part moves from facts to guesses, so treat it as opinion. The near-term shift seems to be from chatbots you ask to agents that act: software that books, files or codes on your behalf with less hand-holding. We've written about agentic AI separately, and it's where most of the big labs are pointing their 2026 releases.
I'd guess that in ten years, "when was AI released to the public?" will sound a bit like asking when electricity was released. The question will still have an answer, but nobody will think of AI as one product with one launch day. Whether that happens smoothly is a different matter; energy costs and regulation could both slow it down. For a wider look, see our piece on the future of AI.
FAQ
Who started AI?
John McCarthy is usually credited with starting AI as a named field. He coined the term "artificial intelligence" in his 1955 proposal and organised the 1956 Dartmouth workshop with Marvin Minsky, Nathaniel Rochester and Claude Shannon. Alan Turing's 1950 paper asking whether machines can think came earlier and shaped the ideas. Our article on who created AI technology has more on each person.
When was AI made public?
Consumer AI became widely public in stages. Siri reached iPhone 4S owners in October 2011, and voice assistants spread from there. The big public moment for generative AI came on 30 November 2022, when OpenAI made ChatGPT free to anyone with a browser as a research preview. Most people mean that date when they ask.
When did AI come out for public use?
For everyday use, the clearest answer is late 2022. ChatGPT launched free on 30 November 2022 and reached about 100 million unique visitors within two months, according to Similarweb figures reported by The Guardian. Claude and Google's Bard (now Gemini) followed in March 2023, so by spring 2023 the public had several chatbots to choose from. Before that, OpenAI's GPT-3 had been available to developers through a waitlisted API since 2020, but you needed some technical know-how to use it. ChatGPT's real change was the chat box anyone could open.
How old is AI?
As a named field, AI turned 70 in 2026, counting from the Dartmouth workshop of 1956. Public chatbots like ChatGPT are just under four years old.
How long has AI been around?
AI has been around as a research field since 1956, so about seven decades. It's been part of everyday consumer products since roughly the 2000s, through search ranking, spam filters and, from 2011, voice assistants like Siri. Our longer explainer on how long AI has been around walks through each decade.
