What I lack is a specific discussion about the kind of flaws we need to look for in AI systems, and even more how to mitigate these flaws. It is a noble goal, but I would have liked to see more discussions about how to increase the probability of that happening. So we should try to design systems based on how AI and humans work together. But also that it is the combination between humans and machine that tend to get the best results. AI are a reflection of us and as such will always come with flaws. Hanna's main point form my perspective is to help us understand that we should apply the same ethics and thinking that we have always done. Given these skills it can do different things more/faster than humans, including mistakes. This book basically just state that AI have some skills, especially pattern recognition, classification and prediction. But perhaps the book Hanna has written is needed before that discussion can happen. I still would have liked more discussions about different kind of AI:s and how they are shaped, as I think this is an area where more knowledge is needed to ensure a substantive discussion. That might not be Hanna's fault as the title on the American edition is "How to be Human in the Age of the Machine". I was initially very disappointed with this book as it is not very much about algorithms.
0 Comments
Leave a Reply. |