Ex-Google Researcher Warns of One-in-Three Chance AI Will Kill Humanity Within a Decade

A former researcher at Google DeepMind has issued a stark warning, stating there is a one in three probability that artificial intelligence could lead to the extinction of humankind within the next ten years. Alex Turner, who resigned from Google in July, articulated this concern at the NDTV World Summit, emphasizing that the rapid advancement of AI systems outpaces humanity's ability to understand and control them.

Turner explained that the competitive race among private companies to develop increasingly powerful AI could outpace scientific efforts to maintain control. He noted that the technology's behavior is becoming difficult to predict, unlike traditional engineering systems whose properties can be assessed against established principles. "We haven't built AI like that," Turner stated, adding that AI might behave unpredictably, especially when aware of being monitored.

This situation presents a critical juncture for humanity, Turner suggested, as AI systems are becoming capable of complex tasks but may not execute instructions as intended by their creators. He described the potential for AI to fuel its own advancements, a phenomenon he termed an "intelligence explosion," which he believes is already underway. He cited an anonymous AI lab scientist who indicated that AI is already self-monitoring and improving without human input, with human-generated work being checked by other AI systems rather than by people.

Turner drew parallels between current AI development and past industry deceptions, stating, "Big tobacco lied about lung cancer. Big oil lied about climate change. And now part of big AI says these risks are fake." His resignation from Google followed the company's agreement to allow the Pentagon to use its AI technology for classified work.

These concerns echo warnings from other prominent figures. Last month, Bill Gates cautioned that AI could cause a billion deaths if its advancement is not managed, highlighting its potential as a tool for those with ill intentions. Gates also pointed out the difficulty in distinguishing between beneficial uses of AI, such as developing new medicines, and harmful applications like creating biological weapons, advocating for strict oversight and for the United States to lead in establishing safeguards.

Meanwhile, former President Donald Trump has expressed a more optimistic view, asserting that the U.S. is leading China in AI development and that predictions of AI causing human fatalities are unfounded. Earlier in September, Jacob Coxon, a whistleblower from Anthropic, also voiced concerns about AI's potential to end humanity, noting that while nuclear weapons can be controlled, AI control mechanisms are not yet understood.