Google says hackers using AI to supercharge attacks

Google revealed on Monday it had thwarted a cyberattack by a criminal group using artificial intelligence to exploit an unpatched vulnerability in a widely used online system administration tool. The attack involved bypassing two-factor authentication, a technique Google described as a zero-day exploit—meaning security teams had no prior warning or time to develop a fix. The company credited AI large language models, likely from outside Google’s Gemini or Anthropic’s Claude Mythos, with helping the hackers identify the flaw. The disrupted operation had no evidence of state sponsorship, though Google noted groups linked to China and North Korea have explored similar AI-driven methods. John Hultquist, chief analyst at Google’s threat intelligence division, warned that criminal hackers now leverage AI’s speed to outpace defenses, increasing risks of data extortion or ransomware attacks. ‘The era of AI-driven vulnerability and exploitation is already here,’ he stated. The incident coincides with broader debates over AI regulation. President Donald Trump’s administration, which reversed Biden-era AI safeguards, announced last week it would evaluate powerful AI models from Google, Microsoft, and Elon Musk’s xAI before public release. However, the policy was later removed from the Commerce Department’s website, reflecting inconsistent messaging. Dean Ball, a former White House tech policy adviser, acknowledged the tension, stating while he opposes excessive regulation, ‘we need to in this case.’ Google confirmed it alerted the affected company and law enforcement, preventing damage. The case underscores how AI tools—originally designed for legitimate purposes—are now being weaponized by cybercriminals to discover and exploit vulnerabilities faster than ever. Experts warn this trend will intensify, forcing industries and governments to adapt their cybersecurity strategies.