We've been calling other things AI since the word was invented. single-layer perceptrons, minimax game solvers, expert systems, any neural networks, video game enemies, keyword-responding text chat, even washing machines (fuzzy logic). Good luck declaring that suddenly it has a new definition.
I wonder if the term changed to distinguish it from the classic non-learning types of AI. You'd be hard pressed to justify that machine learning is not AI while expert systems with hand-written rules are.
> From my experience, having done my bachelor's and master's in this topic, I met no one that used the term AI for anything machine learning.
I'm extremely surprised by this because it's been a very common term for a very long time. You can look back at papers over the last 50+ years and see this. My degree was literally called "Artificial Intelligence and Computer Science" almost 20 years ago, and as far as I can tell it's still called that.
Well but then you'd know how I'd use AI to also encompass ML as ways to achieve AI, wouldn't you? It's a subset but not "AI" in the sense of the word. It's part of AI.
No, not really. I've never heard it used in that specific context (meaning excluding "lower" things) without larger explanation that someone is talking about strong AI or AGI or similar. You can define it like that if you want but you'd be railing against decades of academia and now both users and marketing, and deliberately using terms you expect readers to misunderstand.
AI is a broad field covering far simpler techniques like SVMs even if they're not deemed to be as a way to achieve AI.
Rain is water, even if water covers more than rain. Serverless functions still run on servers.
