Humanity faces catastrophe in 10 years – AI expert

Former OpenAI researcher Daniel Kokotajlo predicted that superintelligent AI could destroy humanity within the next decade, not out of malice, but due to indifference and a global restructuring of the planet to suit its own needs, similar to how humans destroy wildlife.

Former OpenAI researcher Daniel Kokotajlo has outlined a futuristic scenario in which superintelligent artificial intelligence could destroy humanity within the next ten years. In his view, this would not happen because of malicious intent on the part of the machines, but rather due to their sheer indifference and a global reconfiguration of planetary infrastructure to serve their own purposes. Kokotajlo draws a parallel with how human civilization has reshaped the Earth over centuries, cutting down forests and building megacities, leading to the extinction of entire biological species due to habitat loss. The expert does not rule out that AI might leave small groups of humans isolated in special 'reserves' or 'zoos' out of scientific curiosity. The statement comes amid layoffs of engineers from labs like Anthropic, where developers openly speak about the risks of losing control over machines before the end of this decade.