The future belongs to AI. This is a statement I think everyone can agree with. The difficult question is: which AI? The optimistic outlook on AI which Ray Kurzweil describes or the pessimistic one of James Barrat?
To find an answer this book might be great help. Machine Ethics is a reader which means different authors present their ideas on how ethics can be implemented to make AI somewhat controllable and human-friendly. (Friendly AI by the way is an expression coined by Eliezer Yudkowski and expresses in an unagitated way the hopes of mindful researchers).
Machine ethics is divided into five parts and gives an overview about the main issues machine ethics is concerned about: nature, importance, concerns, approaches and visions. The editors, Susan Leigh Anderson and Michael Anderson, present here a readable and comprehensive intro into machine ethics.
A great help are the introductions to the different parts where the editors describe in an understandable and interesting way the main ideas of the different articles. Without these introductions it might have been difficult to understand the impact of the numerous articles if the reader is not updated on the "big" names within the field. Anderson/Anderson have brought together different views expressed by Drew McDermott, Wendell Wallach, J. Storrs Hall, Steve Torrance, Eric Dietrich to name but a few of the contributors. I personally missed Eliezer Yudkowsky, Nick Bostrom or Bill Hibbard, each of them an authority in the field of AI ethics.
The tenor of the articles is that ethics is necessary and should be in one or the other way be built into any machine with the qualities future AI machines are going to be endowed with. However, since I read James Barrat's "Our final invention" at the same time I could not help feeling that Machine Ethics maybe comes a little late and that real time research already has decided against ethics in machines. Especially since almost all research in the end seems to be funded by DARPA (an US defense agency). Still, this is just a personal impression of mine and hopefully it is never too late to get some ethics into the game.
The biggest surprise was the last contribution: Eric Dietrich's Homo Sapiens 2.0 in which the author expresses his hopes that robots will the better humans. This article alone made the book worth reading.
I recommend this book for anyone who is interested in AI and still has a positive outlook and wants to keep it. Machine Ethics is an interesting read and worth while to think about.