Where did AI come from, and why does it feel like it appeared overnight? It did not, and seeing the long road behind it is the fastest way to make sense of both the excitement and the anger pointed at it now.
For most people AI showed up in late 2022, the week a chatbot could suddenly write their emails. One month it was science fiction, the next it was sitting in everyone’s browser. It felt like the technology fell out of the sky fully formed. It did not. AI is old. It has been promising and disappointing us for about seventy years.
It has been letting us down since 1956
The term artificial intelligence was coined back in 1956, by a small group of researchers who were sure thinking machines were only a few years away. They were wrong, and not by a little. The field spent decades overpromising and underdelivering, running into problems far harder than anyone guessed. It got bad enough that people gave the low points a name, the AI winters, long stretches where the money and the faith both dried up and the whole idea looked like a dead end.
What finally made it work
What pulled it out was a quiet change in approach. Instead of trying to program intelligence in by hand, rule by rule, researchers leaned into letting machines learn from data, the idea from the last piece. Slowly that started to work. Around the early 2010s these systems got good at things they had failed at for decades, like recognizing what was in a photo. Then in 2017 came a new way of building the models that handled language far better than anything before it. An old field, finally finding its feet. When one company put that power behind a plain chat box in late 2022 and let anyone use it free, the curtain got pulled back on work that had been building for years. Nothing magical happened that week. It just went public.
I work in communications, and I remember how fast it went from a curiosity to the only thing anyone asked me about. One month nobody around me had heard of it, the next month every meeting found a way to land on it. Nothing about the technology changed in those weeks. The only thing that changed was who could touch it.
You do not need those dates, the shape is what matters. Long slow climb, a couple of real breakthroughs near the end, then a sudden public arrival that made decades of work look like an overnight miracle. That is the whole point.
Hated, and here to stay
That sudden arrival is a big part of why so many people cannot stand it. When something lands overnight and immediately starts touching your work, your art and your feeds, resentment is a fair response. Artists watched models get trained on their work without permission and felt robbed. Writers and designers watched the same tools get aimed at their jobs. Everyone watched the internet fill up with cheap, soulless generated filler. A lot of the anger at AI is not fear of some distant robot. It is a reaction to how it showed up, who profited, and what it is already doing to things people care about. That anger is worth taking seriously.
So AI is neither the miracle nor the menace it gets sold as on a loud day. It is an old idea that finally works, arriving fast, carrying real promise and real cost at the same time. Seeing both of those at once is the whole game. That matters for the next piece, where we get specific about the kinds of AI there are, because AI is not one thing, and the version answering your questions is just one slice of the bigger map.
Next up: The Different Kinds of AI →