Human Compatible: Artificial Intelligence and the Problem of Control (edition 2019)

by Stuart Russell (Author)

Members	Reviews	Popularity	Average rating	Conversations
294	7	89,480	(3.78)	None

fpagan's review

Here's a book from which any reader can get an authentic glimpse of current AI research from a practicing expert in the field. The prospect that true artificial superintelligence may someday be realized raises concerns about what Russell calls the "gorilla problem" (humans' status vis-à-vis robots could become like gorillas' present status vis-à-vis humans!) and the "King Midas problem". He discusses a precautionary approach he calls "beneficial AI", wherein (1) an AI's only objective is the realization of human preferences, (2) those preferences are initially unknown to the AI, and (3) the AI's inferences about the preferences ultimately must come from human behavior. Still needed, he says, are major advances in computational technology ("deep learning", for example, won't suffice) and in moral philosophy (produced, I presume, by philosophers who believe in reaching definite conclusions).

fpagan | May 29, 2020 |

All member reviews

date ▼ | votes

Showing 7 of 7

Important book that covers three main areas:
- a timeline of AI with suggestions on how it will likely evolve and affect us
- an articulation about the risks of AI and what might mitigate them
- an extended discussion that shows how hard it will be to actually implement these controls

The first two areas are extremely well covered and original. The last part of the book is less so and seems more muddled and confusing while trying to advocate for further research in problems which likely do not have a clear solution.

(

)

yates9 | Feb 28, 2024 |

The specific ML recommendations to add uncertainty in the goals is a really cool idea. His policy recommendations are sophomoric and not worth taking seriously at all. (

)

jcvogan1 | Jun 22, 2023 |

It felt like a watered down version of Nick Bolstrom Superintelligence (

)

barrynyle | May 20, 2023 |

Not as good as Brian Christian's The Alignment Problem. (

)

leeinaustin | May 17, 2021 |

Terrible. Sophomoric analysis of fanciful assumptions. It has been too long since someone told the author he is full of it. (

)

jcvogan1 | Sep 13, 2020 |

fpagan | May 29, 2020 |

Excellent book to gain a superficial understanding of the current state of AI. The latter third of the book is a detailed analysis of the role of preferences and utility in AI. This will come more easily to those with some exposure to psychology and behavioural economics, and provides an insight into the complexities involved in getting a machine to do what you want it to do when you are yourself uncertain of what that goal is or what it implies. Great book! (

)