Stuart Russell, professor of Computer Science at UC Berkeley and Director of the Center for Human-Compatible Intelligence (CHAI), has a new book out today: “Human Compatible: Artificial Intelligence and the Control Problem”.
In the book, he explains why he has come to consider his own discipline an existential threat to our species, and lays out how we can change course before it’s too late. The book explores the idea of intelligence in humans and in machines, describes the benefits we can expect (from intelligent personal assistants to vastly accelerated scientific research), and outlines the AI breakthroughs that still have to happen before we reach superhuman AI. Russell spells out the ways humans are already finding to misuse AI, from lethal autonomous weapons to the manipulation of opinions on a global scale. Finally, he explains why we must ensure that we never lose control of machines more powerful than we are. He suggests that we can rebuild AI on a new foundation, according to which machines are designed to be inherently uncertain about the human preferences they are required to satisfy.
I think the book can be an extremely impactful and positive contribution to the field of AI safety. It makes the case for working on AI safety to a broad audience. It also gives updated arguments on AI safety five years after Nick Bostrom’s “Superintelligence” and gives examples of promising research areas. It has also been favorably reviewed by experts in the field, such as Yoshua Bengio and Judea Pearl.
If you would like to support the book’s success, here are some ways to help:
You can now order the book, in the UK edition, and the US edition. The first few weeks seem to be very important in determining a book’s success, so if you plan to read the book eventually (even if you don’t have time now), this is the best time to order!
You can write reviews of the book on Amazon and on other websites, such as Goodreads. Please write genuine reviews. The important thing is to get quite a lot of reviews; even if not every review is favorable, their existence improves the book’s search ranking and makes it more likely that someone studying this topic will eventually stumble across it.
I am working at CHAI, and I’m helping with communication around the book launch. I’d be happy to answer your questions!
Thanks to Aaron Gertler for his help with editing.
Publication of Stuart Russell’s new book on AI safety—reviews needed
Stuart Russell, professor of Computer Science at UC Berkeley and Director of the Center for Human-Compatible Intelligence (CHAI), has a new book out today: “Human Compatible: Artificial Intelligence and the Control Problem”.
In the book, he explains why he has come to consider his own discipline an existential threat to our species, and lays out how we can change course before it’s too late. The book explores the idea of intelligence in humans and in machines, describes the benefits we can expect (from intelligent personal assistants to vastly accelerated scientific research), and outlines the AI breakthroughs that still have to happen before we reach superhuman AI. Russell spells out the ways humans are already finding to misuse AI, from lethal autonomous weapons to the manipulation of opinions on a global scale. Finally, he explains why we must ensure that we never lose control of machines more powerful than we are. He suggests that we can rebuild AI on a new foundation, according to which machines are designed to be inherently uncertain about the human preferences they are required to satisfy.
I think the book can be an extremely impactful and positive contribution to the field of AI safety. It makes the case for working on AI safety to a broad audience. It also gives updated arguments on AI safety five years after Nick Bostrom’s “Superintelligence” and gives examples of promising research areas. It has also been favorably reviewed by experts in the field, such as Yoshua Bengio and Judea Pearl.
If you would like to support the book’s success, here are some ways to help:
You can now order the book, in the UK edition, and the US edition. The first few weeks seem to be very important in determining a book’s success, so if you plan to read the book eventually (even if you don’t have time now), this is the best time to order!
You can write reviews of the book on Amazon and on other websites, such as Goodreads. Please write genuine reviews. The important thing is to get quite a lot of reviews; even if not every review is favorable, their existence improves the book’s search ranking and makes it more likely that someone studying this topic will eventually stumble across it.
I am working at CHAI, and I’m helping with communication around the book launch. I’d be happy to answer your questions!
Thanks to Aaron Gertler for his help with editing.