www.mediafax.ro/life-inedit/un-laureat-nobel-avert

Geoffrey Hinton warns that AI could lead to a major crisis by 2026 or 2027, potentially causing more harm than good. He compares recent AI-related cybersecurity incidents to the Three Mile Island nuclear accident, suggesting a severe event could disrupt essential infrastructures. Hinton emphasizes the need for stricter international regulations and independent testing of AI models, likening it to drug approval processes. While he acknowledges AI's potential benefits, he stresses the importance of managing its risks to prevent significant societal issues.