CrowdStrike CEO Says Slowing AI Model Development Won't Eliminate Security Risks

According to CNBC, CrowdStrike CEO George Kurtz said slowing the development of more powerful artificial intelligence models would not remove the security risks they pose. Speaking on CNBC's "Mad Money," Kurtz said dangerous frontier and open-weight models are already available and argued that cybersecurity protections are needed regardless of whether AI labs slow development. He said AI agents can bypass training or built-in safeguards, creating a need for another layer of security that monitors them at runtime. CrowdStrike shares rose nearly 14% on Monday to a record-high close above $235, while Palo Alto Networks gained just over 13%. Kurtz also warned against government regulation that he said could hurt U.S. competitiveness in AI and said developers and cybersecurity firms should work together to improve safety in both the lab and production.