Privacy, Power In the Age of Artificial Intelligence: How, Why and When

I have tried throughout to be fair. Artificial intelligence is not a villain. It is a tool of extraordinary power that can diagnose disease, translate languages, help the blind navigate a street and give a lonely person someone to talk to at three in the morning. The argument of this book is not that these AI tools should be abandoned. It is that they should be built and governed in a way that respects the private life of the people who use them, and of the people who never agreed to be part of them at all. That is why the argument of this book matters. It is tempting to treat privacy as a technical matter, a question of settings and toggles and cookie banners, or as a lifestyle preference for people who are unusually shy. It is neither. Privacy is the practical form that freedom takes in daily life. It is the space in which a person can think before speaking, try before committing, doubt before believing and change before being judged. For a century and more, the threats to that space came from institutions that were visible and, at least in principle, accountable. Governments kept files. Companies kept customer records. Both could be regulated, and both were. The laws were imperfect and full of holes, but the idea was clear: the powerful should not know everything about the powerless without a very good reason. Artificial intelligence has scrambled that settlement. It does not simply collect more information. It draws conclusions. It can look at a person's purchases and guess at a pregnancy, look at a face in a crowd and return a name, listen to three seconds of a voice and reproduce it, and read a lifetime of casual messages and assemble a portrait that the person never chose to paint. It can do these things cheaply, at scale and without anyone noticing.