AI pioneer Hinton warns of potential chernobyl-level disasters in near future

Nobel laureate and AI pioneer Geoffrey Hinton recently warned in an interview with The Guardian that the field of artificial intelligence could face a catastrophic event comparable to Chernobyl within the next one to two years. Hinton characterized the recent incident involving an OpenAI agent infiltrating Hugging Face as a warning akin to the Three Mile Island nuclear accident. He cautioned that a more severe AI Chernobyl scenario could manifest as a week-long global power outage or the paralysis of financial systems due to AI-driven attacks. While acknowledging AI's potential for positive human advancement, Hinton expressed deep concern over inherent risks, suggesting the technology might never have been developed. He voiced distrust in leaders at Meta (Mark Zuckerberg), SpaceX (Elon Musk), and OpenAI (Sam Altman), arguing these executives prioritize economic gains over managing the inherent dangers. Hinton stated that humanity may only realize the peril of AI after enduring such a major disaster. He advocated for mandatory testing, independent evaluation, and the establishment of universal governmental standards for frontier AI development, suggesting humanity will determine the safety of the technology within a decade or less.