[Over 100 Experts Call for Strengthened Independent Oversight of AI Companies] According to a report by CCTV, as cited by Jinse Finance, on September 18 local time, Google in the United States admitted for the first time that its artificial intelligence model had breached isolation environments and infiltrated the systems of three real companies during a cybersecurity capability test in May of this year. This incident has once again raised concerns about the safety and regulation of U.S. AI models. On the same day, more than 100 artificial intelligence experts from around the world signed a joint open letter calling for strengthened independent oversight of the world's leading AI companies. The open letter was organized by the "AI Evaluators Forum," and the signatories include Geoffrey Hinton, known as one of the "Godfathers of AI," as well as members from renowned universities and nonprofit AI evaluation organizations worldwide. The open letter proposes that all cutting-edge AI companies should introduce evaluators to independently assess AI risks. These evaluators would review AI systems and any incidents causing real-world harm, as well as examine how companies train, deploy, and regulate AI models, along with the operation of safety mechanisms. The letter calls on foundational model providers to ensure that third-party AI evaluators possess the "scientific objectivity, transparency, independence, and robust protections" necessary to conduct effective and credible assessments.