In the discourse of AI safety, there is a gap in the conversation, and this book aims to fill it. While most argue over the technical problem of alignment, no one really delves into the systemic issue with the necessary depth. That is: even if we can align superintelligence, how do we force competitive actors to meaningfully deploy it on a global scale? When winning is everything and safety is a distant priority, how can corporations, countries, and labs in competition risk handicapping their ASI with barriers to optimisation if it means losing the race?

While most critics only touch on these arguments Driven to Extinction dives deep into the logic of competitive dynamics and game theory when applied to an intelligence so vast it can effectively rewrite its own reality.

This is not a conversation most are ready for. There is no hedging language that frames the majority of AI safety discussions. It's just rigorously established premises leading to their most logical conclusions - and those conclusions are as robust as they are dire.