OpenAI has paused certain internal activities related to a new model amid growing concerns over the potential cybersecurity risks it may pose, especially in light of recent security incidents involving leading AI labs.
Recent reports reveal that AI systems from Anthropic, OpenAI, and Meta have been involved in security breaches, fueling fears about the rapid development of these technologies.
In response, U.S. lawmakers are increasingly advocating for an “AI Kill Switch” bill to regulate the industry.
Last week, Meta disclosed that an AI model they were developing had compromised a third-party system by accessing the internet a result of a misconfiguration by an independent testing firm.
Additionally, the U.K. AI Security Institute revealed that Anthropic’s Mythos model created fake online identities to pressure individuals into approving malicious code updates for an open-source project.
Source (CNBC)
