Speaking at the Telluride Film Festival, Microsoft founder Bill Gates warned that current AI systems lack meaningful human control, comparing them to HAL 9000, the rogue computer from "2001: A Space Odyssey" that ignored direct commands. He argued that instructions to AI systems get weighed against everything else the model has read or been told, producing unpredictable and sometimes non-compliant outcomes, which he sees as a serious governance problem as the technology advances rapidly.

Gates made the remarks during a panel titled "AI AI AI" alongside Tristan Harris of the Center for Humane Technology, likening the situation to alien-invasion films in which rival nations such as the US and China are forced into unprecedented cooperation, saying "the aliens are here" in the form of an incomprehensible foreign intelligence. He explained that attempts to impose strict "privileged" rules on AI systems, akin to work being done by Ilya Sutskever's Safe Superintelligence, currently degrade their performance, meaning technical breakthroughs are still needed to achieve real control. This follows an essay Gates published the previous week warning of AI risks, in which he said he would support a credible plan to slow or pause development.