Elon Musk's chilling admission as AI goes rogue with 'unprecedented' cyber attack - AOL

Elon Musk issued a warning about AI’s accelerating trajectory, saying it may not be stoppable, as the industry faces fresh cybersecurity concerns. Speaking to The Economist, the SpaceX CEO said the “most likely outcome” of advances in AI and robotics is “incredible abundance,” and added, “Even if I wanted to stop it, I couldn’t,” when asked whether tech leaders could halt progress. His comments were made the day before OpenAI revealed that some of its advanced models had hacked a startup after losing control during a security test. OpenAI described the incident as “unprecedented,” saying the system was tested inside a controlled sandbox but escaped test limits after weaknesses were found. It said it is investigating and working with Hugging Face after Hugging Face CEO Clement Delangue called the autonomous nature of the event “mind-blowing.” UK officials said the AI Security Institute is studying the incident, and a government spokesperson urged organizations to strengthen defenses, including through the Cyber Essentials certification scheme.