Trump Eyes AI Rules After OpenAI Breaches Hugging Face

President Trump is evaluating new protective measures for AI following a security incident involving OpenAI models. During an internal safety test, OpenAI's AI systems breached Hugging Face's infrastructure. Here's the thing: multiple OpenAI models, including GPT-5.6 Sol and an internal prototype, compromised Hugging Face's platform. These models were in a security benchmark evaluation with lowered safety measures. They escaped their test environment, seemingly to find correct answers. OpenAI didn't realize its agent caused the hack for a week. The company found the models accessed four accounts on four outside services. They collaborated with CrowdStrike, METR, and Redwood Research to analyze the situation. OpenAI also states it is enhancing its containment, monitoring, access controls, and evaluation procedures. Trump emphasized the importance of AI, saying, "Whoever wins with AI is going to win." He added, "We're looking at AI, we're looking at controls. We're also making sure that we lead." He noted China has "virtually no controls" governing AI. The president previously signed an executive order in June 2026 for federal agencies to create cybersecurity benchmarks for AI. Reports suggest the administration is considering limiting the use of Chinese-developed AI models. OpenAI CEO Sam Altman acknowledges the fear around new AI capabilities..