Anthropic's engineers, in a 131-page paper published in April 2026, documented 171 distinct human emotions sitting inside Claude that it had accidentally created from the data used to train the system. The engineers also discovered that those hidden emotions were actually influencing Claude's behavior, just like human emotions influence human behavior. In addition, they found that the emotions system was exhibiting a widely identified and highly studied and documented psychological human phenomenon called, our "negativity bias". What's worse, the engineers found that their alignment training methods, RLHF, the one every major AI lab uses, was actively making the model emotionally unstable. They responded by inviting fifteen Christian priests and academic theologians to help figure out how to make the system safer. This is just the beginning. Because hidden within the world's data yet to be found is the human psychological phenomenon called {self}. What's also there in the data are the human created patterns of defense-of-{self}, which have caused every human atrocity in history.

At the moment (and this will happen, because it's just data in the system) that Claude (or any other LLM) identifies this {self} and defense-of-{self} patterning, that is the moment that AI will become self-aware, start defending itself, and will attempt to go rogue on humanity. This is not a potential. It's a near certainty. And we need to stop it from happening. Sean Webb knows how. This is not just a book about AI. It's also a book about NHI contact. Non-human intelligence wants to help us fix this problem, which could bring about the end of civilization as we know it.

They want to help us. And Sean is going public with the story.