THE DANGER WAS NEVER ARTIFICIAL.
A mother answered the phone to a fake of her daughter's voice. A teenager confided in a chatbot at two in the morning. A finance employee joined a video call with colleagues who did not exist. A town watched its taps run dry. A government resigned over a biased algorithm.
There are already plenty of books about frameworks. This one is about people.
I advise organisations on AI governance, and for a little over a year I read incident reports most days. What I could not get past was how often nobody involved understood what they had deployed, and how little that seemed to bother anyone.
The risks are already documented but they are just scattered. A regulator sees the enforcement side, a lawyer sees the litigation, a company sees its own incident, and a parent sees a chatbot their child talks to at midnight. Almost nobody sees all of it, which is why the same failure keeps happening at every level.
So I built the map. Nine circles, sorted by severity. Limbo, Lust, Gluttony, Greed, Wrath, Heresy, Violence, Fraud, Treachery. From the ordinary failure of not understanding what a system does, down to whether the most capable systems can be trusted to do what we meant when nobody is watching.
Every case is documented, drawn from courts, regulators, company disclosures and published research. All the sources are listed chapter by chapter, so you can check any claim yourself. Where a company disputed an account, its response sits in the same chapter as the criticism. The book separates what has already happened from what could, and says which is which on every page.
Then it argues that this is governable, because humanity has governed dangerous power before.
Written for people who have to make decisions about these systems, in compliance, legal, policy, technology and risk, and for anyone living alongside them.
If you are not sure this is for you, that is where the book starts.