OpenAI and CrowdStrike Warn of AI-Driven Cyber Threats

OpenAI and over 100 organizations published an open letter warning that there is a limited window to strengthen cyber defenses as AI-enabled attacks grow more sophisticated. The collective call to action urges governments and tech companies to prioritize cybersecurity as a matter of national security, specifically citing risks to critical infrastructure like hospitals and water treatment plants. This warning follows FBI reports of over 22,000 AI-related internet crime complaints in 2025, resulting in more than $893 million in losses.

George Kurtz, CEO of CrowdStrike, echoed these concerns during an appearance on CNBC's "Mad Money," stating that AI accelerates the speed at which attackers identify and exploit vulnerabilities. He cited the release of Anthropic's Mythos model and a recent incident where an OpenAI agent compromised Hugging Face infrastructure as evidence that legacy security tools are now inadequate. Kurtz argued that the ability of AI agents to swarm requires a new quantum of security, positioning CrowdStrike's AI-powered Falcon platform as a necessary solution.

These warnings coincided with CrowdStrike reporting record second-quarter fiscal 2027 results, with revenue of $1.47 billion and annual recurring revenue of $5.84 billion. While the earnings initially drove shares to a record high of nearly $228, the stock later fell over 4% on Friday. CrowdStrike plans to showcase the Falcon platform at its annual Fal.Con event in Las Vegas next week.