As If Human: Ethics and Artificial Intelligence

by Nigel Shadbolt, Roger Hampson

A new approach to the challenges surrounding artificial intelligence that argues for assessing AI actions as if they came from a human being
 
Every day we are confronted with ethical challenges arising from machine-mediated decision-making. Is the facial recognition software used by an agency fair? When algorithms determine questions of justice, finance, health, and defense, are the decisions proportionate, equitable, transparent, and accountable? How do we harness technology to empower rather than oppress?
 
Nigel Shadbolt and Roger Hampson argue that we can avert a disastrous and amoral future by assessing artificial-intelligence decisions, products, and calls for action as if they came from a human being. The authors go beyond the headlines about rampant robots to apply established moral principles, setting out seven “proverbs” for shaping this
 
● A thing should say what it is and be what it says.
● Artificial intelligences should show respect for human beings.
● Artificial intelligences are ethical only if they embody the best human values.
● Artificial intelligences should be transparent and accountable to humans.
● Humans have a right to be judged by humans if they so wish.
● Decisions that affect a lot of humans should involve a lot of humans.
● The future of humanity must not be decided in private by vested interests.

Genres: Nonfiction, Artificial Intelligence, Philosophy

272 pages, Hardcover

Published May 14, 2024

Reviews

[Sepehr · Rating 1 out of 5]

The authors, half-way through the book (on ethics and AI no less) ask “Does artificial intelligence raise new ethical challenges?”

[Lucy Snow · Rating 3 out of 5]

3.5 stars

I really wanted to love this book as it covers quite an interesting topic and grounds itself in philosophy much more than other books I’ve read on the issue. Unfortunately the prose was quite difficult to get through, I was unsure what the authors were trying to say at times even when rereading the paragraph a couple times. It also requires the reader to already know some philosophical concepts as it makes references to some without explaining what it means (or sometimes even how it’s relevant to the discussion in the book). Similarly it mentioned people’s names (such as some animal rights person) seemingly for no reason other than to just mention them. Personally, as someone not from England and is not familiar with these people, I found this practice confusing — if you have a point to make then just make it, don’t give me homework to google someone and understand why they are relevant.

Continuing this line of thought, I felt like the arguments actually discussed were often left without a proper conclusion or enough explanation, such as the authors having an issue with claims that machines cannot be viewed the same as humans when they infringe privacy because it’s a giant “as if” argument, all the while arguing that the two should be viewed the same in moral arguments. That would be fine (all analogies and examples can have nuance) if they would actually explain why the “as if” argument does not work here specifically, instead I was left to figure out why by myself, guessing why the authors thought so. Similarly at some point the authors say that they have objections to the mass collection of data but fail so say why beyond stating that it’s not connected to privacy.

Overall while I found the arguments presented engaging, I’m just struggling to understand the purpose and the target audience of this book. It is certainly not an introductory book into the field of AI ethics or philosophy, but I also don’t see how it contributes to the body of works on the topic enough to recommend it to someone over some other book.

Maybe once I read more on philosophy I’ll revisit this book and find my opinion changed but for now 3.5 stars.

[Jeremy · Rating 5 out of 5]

An important book that raised all sorts of issues that I had not considered in the context of AI. This really helped me work my way through the thinking.

[Aki Härmä · Rating 3 out of 5]

I didn't know about the authors before buying the book but was interested in the description of the book, mostly because it is a very new (post-chatGPT, that is) book about the ethics and AI. I am not sure I understood their big statement that machines should be judged "as if" they were human. I have a feeling that the authors also started to lose the grip of the idea around midway. The book has interesting content and ideas. One idea relate to artificial unfairness. If a machine learning algorithm was trained using biased data, it makes unfair decisions. We do not think unfairness of a computer as "artificial unfairness" but as "real unfairness." I can buy this, and actually used it in a talk earlier this week, but, I do not see it really extending into all other virtues. Maybe we can say that an algorithm has "real intelligence" or "real compassion", if a human observing the machine feels that way. Clifford Nass has written a lot about us projecting human attributes to machines already 30 years ago and now Shadbolt & Hamson say this is what we actually should do. Towards the end of the book the authors started describing their own research projects, which was frankly not very interesting and also drifted the book even further away from any conclusions.

[سليمان العوشن · Rating 3 out of 5]

عند قراءتك لهذا الملخص، تشعر أنك أمام كتاب لا يكتفي بشرح الذكاء الاصطناعي بوصفه تقنية، بل يدعوك للتفكير فيه كقوة أخلاقية مؤثرة في مصائر البشر. ينجح الكتاب في أن يضعك مباشرة أمام السؤال الجوهري: إذا كانت قرارات الذكاء الاصطناعي تؤثر في حياتك كما يفعل الإنسان، فلماذا لا تُحاسَب أخلاقيًا بالطريقة نفسها؟
أنت تلمس في الطرح توازنًا واضحًا بين العمق الفلسفي والواقعية التطبيقية، خاصة عند مناقشة مبادئ العدالة والشفافية والمساءلة. أكثر ما يلفت انتباهك هو تركيز المؤلفين على المجالات الحساسة مثل القضاء والطب، حيث لا يكون الخطأ مجرد خلل تقني، بل مسألة تمس كرامة الإنسان وحقوقه
هذا الكتاب يجعلك أكثر وعيًا بخطورة تجاهل البعد الأخلاقي، ويمنحك إطارًا فكريًا يساعدك على تقييم الذكاء الاصطناعي بعين ناقدة ومسؤولة، لا منبهرة ولا رافضة


تابعني على شبكات التواصل الاجتماعي

سليمان العوشن

http://about.me/aloshan

[Mercedes · Rating 1 out of 5]

This book could've easily been a paper; it's reiterative and overall, boring. I found some parts hard to read, as if random sentences (and chapters) were written separately and just stitched together in the end. Ironically it had me thinking at times if it had been written by an AI.
Specifically, the whole chapter on privacy had me grinding my teeth. And don't get me started on all the talk about Microsoft (is this some sort of propaganda for the company?).
Towards the end, it goes more and more off-topic, which leads me to my first point: are they just trying to fill pages so they can have a decent length book?
Even the conclusions seem lazy, just a bunch of well wishes; that's definitely not worth my time.