On April 11, CNBC reported that, ahead of AI company Anthropic's launch of its new model, Mythos, US Vice President Vance and Treasury Secretary Bessent organized a video conference last week. The virtual gathering brought together the CEOs of prominent tech firms, including Anthropic, xAI, Google, OpenAI, and Microsoft. The primary focus of the meeting was to delve into the safety aspects of AI models and to devise strategies for countering cyberattacks.

During the conference, participants engaged in in-depth discussions regarding the present safety landscape of large language models, deployment standards, and policy safeguards. US officials voiced their apprehensions about the potential for these models to be manipulated by malicious hackers. Anthropic, for its part, has provided the government with an overview of Mythos's capabilities and is actively collaborating on safety assessments. Meanwhile, responses from other companies are still pending.