Palo Alto warns AI-driven cyberattacks could become routine within months

Palo Alto Networks Chief Technology Officer Lee Klarich warned that companies may have only three to five months to strengthen software defenses before AI-driven vulnerability attacks become routine, CNBC reported, citing a May 13, 2026, blog post. Klarich said increasingly capable AI models are improving hackers’ ability to find software flaws, raising pressure on corporate security teams to prepare for large-scale attacks using unknown vulnerabilities. Google parent Alphabet (GOOGL) said this week it blocked an attempted AI-enabled mass exploitation event. Klarich called for industrywide development of tools that can proactively detect new attack methods, including virtual patching. He said Palo Alto (PANW) plans to release its first related capabilities soon. Anthropic limited access in April 2026 to its Mythos model for testing by selected companies, including Palo Alto, CrowdStrike (CRWD), Amazon (AMZN), Apple (AAPL) and JPMorgan Chase (JPM). OpenAI also recently introduced GPT-5.5-Cyber and its Daybreak cybersecurity program.