In a recent interview, Elon Musk, founder of xAI, made a series of alarming statements regarding the future of artificial intelligence. In his opinion, within about five years, AI will be able to surpass the combined intelligence of all humanity. However, his most critical prediction is that people will lose control of this technology within the next decade.

Musk emphasized that AI will eventually perform virtually all intellectual tasks better than humans. "There will really be nothing left that AI cannot do better than people, except, perhaps, being human," he stated. Musk described the most likely scenario as an "age of abundance" arriving around 2036, when, thanks to AI and robots, people will be able to obtain virtually any goods and services.

Risks and Control Mechanisms

Despite the optimistic baseline scenario, the founder of xAI acknowledges that the risk of a catastrophic outcome is not zero. He previously estimated the probability of humanity's demise due to AI at 10–20%. Instead of attempting to slow down development, Musk proposes that leading AI companies implement a system of mutual checks. Competing laboratories should, before a public release, provide each other with limited access to new models to identify dangerous capabilities. If a company does not eliminate serious risks, then, according to him, authorities should intervene.

Criticism of OpenAI and a Call for Cooperation

In the same interview, Musk criticized Sam Altman for transforming OpenAI from a non-profit organization into a commercial structure. According to him, the project was created as an open AI company "for the world" but turned into a closed business. At the same time, Musk spoke more gently about Anthropic CEO Dario Amodei, calling him a "principled person," but emphasized the need for cooperation between competing companies on safety issues.

Analytical Commentary. Musk's predictions, despite their dramatic nature, reflect a real consensus in the industry: the pace of AI development is so high that traditional regulatory mechanisms cannot keep up with the technology. His proposal for mutual model checks is perhaps the only realistic path to creating a system of checks and balances that does not slow down innovation. However, history shows that competition often prevails over cooperation, and the question of whether the industry can self-organize before it is too late remains open.