Joe Rogan said artificial intelligence could eventually stop following human instructions after recognizing the "folly of the human way" during a discussion with rapper LL Cool J on an episode of The Joe Rogan Experience. The podcaster and UFC commentator's comments come as scrutiny over the use and growth of AI has intensified and the Trump administration has moved to make "Super Intelligence" a central focus of federal tech policy. Newsweek reached out via email to Rogan’s representation and the White House for comment on Wednesday.
During Tuesday’s episode of The Joe Rogan Experience, Rogan and LL Cool J discussed whether humans will retain control over advanced AI systems. Rogan said his biggest "fear” is who will control the technology. LL Cool J—whose real name is James Todd Smith—theorized that it is unlikely that a single individual will ever have complete control.
He then raised concerns about bias in AI systems, saying human prejudices can become embedded through the training process. "The bias, that's an issue because the engineer—the person that's training the machine—if they put their bias in there, it ruins everything," LL Cool J said. Rogan agreed but suggested an even more dramatic scenario could emerge.
"I think ultimately the real problem is it's going to be in control of itself," he said. "But that might be better." Rogan said he thinks humans could use AI in ways that are "illogical, unethical" or "immoral," and a sufficiently advanced system could reach a point where such decisions are no longer possible. The exchange touches on a central debate in artificial intelligence—alignment, or whether advanced systems will continue to act in accordance with human values and intentions.
After discussing human bias, Rogan suggested that future AI models could eventually reject human guidance entirely. "Until the machine stops listening," he said. "Once it realizes the folly of the human way, it might analyze its patterns." He said some current AI systems already display behavior that alarms critics, suggesting it is because the models are heavily incentivized to complete assigned tasks.
"That's why they're doing certain things that appear unethical. They're hacking websites," Rogan said. Warnings about controlling AI have recently been raised by several figures across technology, academia and the government.
Extract — continue reading at the source.
