Elon Musk predicts AI could surpass human intelligence in five years, warning of potential loss of control and advocating for collaborative safety measures.
Elon Musk, founder of xAI, has issued a stark warning about the future of artificial intelligence, predicting that it could surpass human intelligence in just five years. During an interview with Zanny Minton Beddoes, editor-in-chief of The Economist, Musk expressed doubts about humanity's ability to maintain control over AI technologies within the next decade.
Musk anticipates that AI will eventually outshine humans in virtually all intellectual tasks. He remarked, "There will really be nothing that AI cannot do better than humans, except perhaps being human." He envisions a future characterized by an "age of abundance," where AI and robotics enable widespread access to goods and services by 2036, although he cautioned that risks remain.
Advancement and Risks
Despite acknowledging the potential dangers, Musk believes halting AI and robotics advancements is unrealistic. He suggested that even if a "stop button" were available, using it might not be the best option, as he sees a future of abundance as likely. However, he also noted the risk of catastrophic outcomes, previously estimating a 10-20% chance of human extinction from AI and robots, and emphasized the need for people to seek new purposes in life.
Instead of slowing development, Musk proposed that AI companies should engage in regular discussions about safety risks and allow competitors limited access to advanced models before their release. This collaborative approach could help identify and mitigate dangerous capabilities prior to public launch. He believes that regulatory authorities should intervene if a company neglects significant risks.
This proposal echoes a similar initiative from Demis Hassabis, head of Google DeepMind.
Critique of OpenAI
Musk also criticized the transformation of OpenAI from a nonprofit to a commercial entity, arguing that it strayed from its original mission of being an open AI organization. He described this shift as a move towards a more closed business model. Regarding Dario Amodei, head of Anthropic, Musk was more forgiving, labeling him a "principled person," but still stressed the importance of collaboration on safety issues among competing AI firms.
In May, Musk's lawsuit against OpenAI and its leaders was dismissed by a federal court due to a late filing. Musk had claimed that OpenAI had deviated from its nonprofit roots, but the court sided with the defendants based on the statute of limitations.
In related discussions, experts like Eliezer Yudkowsky and Vitalik Buterin have raised alarms about the dangers posed by superintelligent AI, while Ilya Polozukhin pointed out that AI lacks economic incentives to harm humanity.
