Musk warns AI may outstrip humanity in five years

Elon Musk warned in a recent interview with an editor that artificial intelligence could exceed the combined intelligence of humanity in about five years and that humans may lose control of the technology within a decade.

He predicted an era of “amazing abundance” in which technology could produce almost anything people imagine, and he warned that the same progress could create an existential risk if not managed.

Musk said he did not see a realistic way to halt AI development, adding, “I can’t see any way to really stop this incredible momentum of AI and robots.” He noted that pausing development might be unwise because the technology could bring significant benefits.

He proposed that the world’s leading AI developers hold regular meetings to discuss security risks and that competing labs receive limited early access to each other’s frontier models so teams can test for hidden dangers before public release. He suggested governments should have authority to intervene if companies fail to address serious safety problems identified through such cooperation.

Musk criticized OpenAI’s shift from a nonprofit to a for-profit company and faulted its leadership for departing from the organization’s original mission. In May, a California jury rejected his $150 billion lawsuit against OpenAI and its leaders, finding they were not liable for claims that they abandoned the nonprofit mission.

At the same time he praised Anthropic’s co-founder Dario Amodei as principled and described employees he has met there as well-intentioned. He acknowledged concerns that some companies use fear-based messages to promote products and warned against complacency, saying “the road to hell is mostly paved with bad intentions.”

Musk framed his safety proposals as practical steps rival teams could take despite personal animosities, adding, “At the end of the day, if we have to talk, we’ll talk. Set aside our personal differences for the good of the world.”

Musk founded xAI to pursue AI development and has repeatedly urged more coordination among leading labs and possible government oversight to manage risks from rapid progress.