Vice President JD Vance and Treasury Secretary Scott Bessent recently questioned leading technology CEOs regarding the security posture of artificial intelligence models, particularly ahead of Anthropic's Mythos release.
High-Level Security Review of AI Models
According to sources familiar with the matter, the private meeting, which took place over the phone, focused on assessing the safety and resilience of Large Language Models (LLMs). The discussion was prompted by the anticipation of Anthropic launching its new Mythos model.
Participants in the high-level discussion included key figures from major tech corporations and cybersecurity firms:
- Government Officials: Vice President JD Vance and Treasury Secretary Scott Bessent.
- Tech Leaders: Sundar Pichai (Google), Sam Altman (OpenAI), Satya Nadella (Microsoft), and Dario Amodei (Anthropic).
- Security Experts: George Kurtz (CrowdStrike) and Nikesh Arora (Palo Alto Networks).
Core Focus: Mitigating AI Risks
The primary objective of the meeting was to establish best practices for the safe deployment of advanced AI. Officials specifically addressed:
- Security Posture: Evaluating the inherent security vulnerabilities of LLMs.
- Safe Deployment: Discussing methods to ensure models are used responsibly and securely.
- Threat Response: Developing strategies to respond if AI models are scaled or misused by malicious actors.
Industry Response and Cooperation
While OpenAI and Anthropic declined to provide detailed comments on the meeting, both companies confirmed their engagement with federal authorities regarding cybersecurity.
- Anthropic: Confirmed that the company has been in contact with White House officials in recent weeks and stated its readiness to support the government's own testing and evaluation of the technology.
- OpenAI: Declined to comment on the specifics of the meeting.
CNBC reached out to the White House and the participating companies for further comment.
