AI-driven cyberattacks will be the ‘new norm’ in months
As the ransomware industry evolves, experts predict that hackers will continue to find more and more ways to use technology to exploit businesses and individuals.
Sexy Mongkhonkhamsao | An | Getty Images
Palo Alto Networks tech chief Lee Klarich says companies are wasting time taking action software defenses Hackers are increasingly exploiting vulnerabilities with the help of artificial intelligence models.
“We now estimate a narrow window of three to five months for organizations to outpace competitors before AI-driven exploits become the new norm,” he wrote in a blog post Wednesday. “This coming deluge of vulnerabilities demands urgency.”
The rise of increasingly complex AI models like Anthropic’s Mythos has increased the risks and put pressure on cybersecurity teams to ramp up their defenses as they prepare for a wave of cyberattacks that could exploit previously unknown software vulnerabilities. Concerns led the White House to hold meetings with bank leaders and tech giants.
Google This week, it was announced that it had halted an attempt to use AI for a “mass exploit incident,” but hackers are already using existing AI tools to exploit software vulnerabilities.
Klarich acknowledged that these features would not be limited to newer models and called for industry-wide innovation to catch up with new attack techniques, including virtual patching capabilities. He said Palo will be rolling out its first set of capabilities “very soon.”
Last month, Anthropic limited the rollout of Mythos to a select group of companies to test and fix vulnerabilities before they could be exploited by hackers. The group included Palo Alto Networks. CrowdStrike, Amazon, Apple And JPMorgan.
OpenAI announced the GPT-5.5-Cyber model last week, and it Dawn cyber enterprise.
“The big question a few weeks ago was: ‘Are we overestimating model talent?’ With more testing, I can safely say that we are not,” Klarich wrote. “In fact, these models are probably better at finding vulnerabilities than we initially realized.”
