Overview
- Musk, speaking to The Economist on Friday, said artificial intelligence could surpass the combined intelligence of all humans in about five years and that people are unlikely to control it within a decade.
- He said AI progress is effectively unstoppable, that the risk of catastrophic outcomes is not zero, and that the most likely near-term result is an era of broad abundance rather than guaranteed disaster.
- To reduce danger, Musk proposed regular industry safety talks and weekly or biweekly peer reviews where leading labs examine each other’s frontier models before public release.
- He acknowledged his own role in accelerating AI by helping to found OpenAI, criticized OpenAI’s shift toward a closed commercial model, and praised Anthropic’s leadership under Dario Amodei.
- The comments sharpen a debate over whether self‑coordination among rival companies can curb existential risks and signal possible pressure on labs and regulators to adopt formal model‑review and disclosure practices.