What you’re saying expressly isn’t true. Academically, deep learning is considered a subset of machine learning is considered a subset of artificial intelligence.
- Deep learning is machine learning that makes use of deep neural networks.
- Machine learning is artificial intelligence which can perform tasks without explicit instructions by learning from a dataset and generalizing to other data.
- Artificial intelligence is simply trying to make a computer display some sort of intelligence that’s seen as human-like. For example, a perceptron is artificial intelligence because how could a computer possibly see like a human? Chess bots are artificial intelligence because it was thought that chess represented some sort of higher intelligence unique to humans. NPC actions in video games can be artificial intelligence because you’re simulating what another human might do.
Would you like the textbooks from 10 years ago on this exact subject that I’m referencing? The term AI hasn’t been co-opted; you might’ve simply been thinking of general artificial intelligence, because “pretty much any form of machine learning” has been called AI since the dawn of machine learning – because it is.
That toaster is what AI is. If it’s machine learning, it’s AI. If I make a toilet that uses a shitty-ass single-layer perceptron to decide when to flush, that’s an AI-powered toilet even if it’s a worthless piece of crap. You can be disenchanted with it as a gimmick all you want (I am too), but it falls under AI the same way it has since the 1950s. The marketing way of referring to things you just showed me entirely comports with the academic one provided what the label says is true.