This video features an interview with Jeffrey Hinton, the Nobel Prize-winning “Godfather of AI,” where he expresses his significant concerns about the rapid development of artificial intelligence.
Hinton’s Warning
Hinton states that while the rapid progress of AI could solve problems like climate change, it primarily makes him “really worry”. He compares the situation to raising a “really cute tiger cub”, warning that unless one can be “very sure that it’s not going to want to kill you when it’s grown up, you should worry”.
He believes people have not yet understood what is coming, and he estimates a 10% to 20% risk that AI will ultimately “take over from humans”. The challenge, he says, is designing AI in a way that it “never wants to take control” and is “always benevolent”.
Key Concerns and Corporate Behavior
- Societal Harm: Hinton predicts AI will make authoritarians “more oppressive” and hackers “more effective”. He is so concerned that he has spread his own money across three banks.
- Speed of Development: Hinton admits that while he believed the technology was possible, he didn’t think the industry would get to this point “not this soon”.
- Lack of Regulation: Hinton is critical of major tech companies, stating they are “foolishly, selfishly putting all of humanity at risk” by racing against each other. He notes they are lobbying for less AI regulation to protect “short-term profits”. He argues that companies should spend a “significant fraction like a third” of their computer time on safety research, but “right now it’s much much less”.
- Military Use: He expressed being “very disappointed” when his former employer, Google, went back on its promise not to support military uses of AI.
Hinton’s foundational work in 1986, which proposed using a neural network to predict the next word, is the concept that today’s large language models are built upon.
The full video is available here: “Godfather of AI” has new warning about artificial intelligence: “You should worry”