Artificial intelligence seems to be the go-to solution to every problem there is (technological in nature or otherwise), and it’s only getting worse. A staggering number of both startups and established companies are loudly proclaiming how AI, or machine learning, or deep learning, or whatever is absolutely going to make everything faster, better, cheaper, fairer, and so on. And there’s a lot of so on.

The reason that this sort of breathless and inevitably shallow media-driven enthusiasm for artificial intelligence is effective is because there’s just enough of a general understanding of AI for people to know that it can do some cool things, but not so much of an understanding for people to question what it’s actually capable of, or whether applying to to a specific problem is a good idea. This is not to say that a lack of understanding is anyone’s fault, really: it’s hard to define what AI even is, much less communicate how it works. And without the proper context, there’s no way to make an informed judgement about the future potential of artificial intelligence.

In his new book, Architects of Intelligence: The Truth About AI from the People Building It, Martin Ford somehow managed to interview 23 of the most experienced AI and robotics researchers in the world, asking them about the current state of AI, how AI can be applied to solve useful problems, and what that means for the future of robotics and computing.