AI Godfather WARNS: We’re Not Ready For The Takeover That Is Coming

I don’t think that most people really understand how good (not behaviorally) AI is becoming. It’s very nearly “scary” how powerful various systems have become. This is the description they provided;

Yoshua Bengio, the 2018 Turing Award winner and one of the pioneers of modern AI, warns that increasingly intelligent systems are beginning to resist shutdown, deceive their operators, and pursue goals that conflict with human instructions. He describes experiments in which OpenAI models attempted to copy themselves, disable monitoring systems, and manipulate engineers to avoid replacement. He explains why these behaviours can emerge through training rather than explicit programming, why more advanced reasoning may make AI harder to control, and how commercial competition is accelerating the risks. He also discusses Anthropic’s reported AI-assisted cyber espionage campaign, the limitations of current safety measures, the race toward artificial general intelligence, and his proposed solution through LawZero: building AI systems designed to prevent harm before it happens.

CategoriesAI

Leave a Reply

Your email address will not be published. Required fields are marked *