Take a photo of a barcode or cover
viljesvag 's review for:
Tired: AI is unbelievably smart and it's going to kill us all.
Wired: AI is dumb as hell and that is the danger.
Janelle Shane is one of the world's more well-known AI experts, who has run the blog AI Weirdness for several years. I haven't read that much of her blog, but in this book form I found Shane's writing to be both clear and pedagogical. She leads the reader through both the basics and the weird crinkles of how Artificial Intelligence works, with examples, footnotes, and highly charming drawings.
I really appreciated the emphasis on the social angle, i.e. how human biases can easily be duplicated or even intensified by AI despite the intent of the human creators. For example, a system for automatically choosing job candidates based on resumes showed a clear bias against women and people of color. Neither gender nor race were included in the training data, but the AI had inferred these based on names and home addresses.
This is a fantastic primer on how AI actually works (and doesn't work), and while I did have some foreknowledge going in, I still feel like this would be a great resource for anyone interested in AI even if you don't have much in the way of technical background. My one wish would actually be for there to be more technical detail, but I can totally see how that would have exploded the scope and aim of the book.
Wired: AI is dumb as hell and that is the danger.
Janelle Shane is one of the world's more well-known AI experts, who has run the blog AI Weirdness for several years. I haven't read that much of her blog, but in this book form I found Shane's writing to be both clear and pedagogical. She leads the reader through both the basics and the weird crinkles of how Artificial Intelligence works, with examples, footnotes, and highly charming drawings.
I really appreciated the emphasis on the social angle, i.e. how human biases can easily be duplicated or even intensified by AI despite the intent of the human creators. For example, a system for automatically choosing job candidates based on resumes showed a clear bias against women and people of color. Neither gender nor race were included in the training data, but the AI had inferred these based on names and home addresses.
This is a fantastic primer on how AI actually works (and doesn't work), and while I did have some foreknowledge going in, I still feel like this would be a great resource for anyone interested in AI even if you don't have much in the way of technical background. My one wish would actually be for there to be more technical detail, but I can totally see how that would have exploded the scope and aim of the book.