AI-powered hacking has exploded into industrial-scale threat, Google says

Google reports that AI-powered hacking has rapidly escalated into an industrial-scale threat within three months. State-linked actors from China, North Korea, and Russia, along with criminal groups, are reportedly using commercial AI models like Gemini, Claude, and OpenAI tools to enhance the speed, scale, and sophistication of cyberattacks. This includes refining operations, building better malware, and exploiting vulnerabilities. The report highlights that threat actors are already leveraging AI for these purposes, moving beyond theoretical future risks. Anthropic previously withheld a powerful AI model due to its potential to find zero-day vulnerabilities, underscoring the growing capabilities of AI in cybersecurity, both for offense and defense.

China is accelerating its development of AI-driven cybersecurity solutions in response to advancements by US companies like Anthropic and OpenAI. Anthropic's new model, Mythos, launched in April, has demonstrated a significant ability to identify and exploit cybersecurity vulnerabilities quickly. While Chinese cybersecurity AI models currently lag behind those of their US counterparts, their capabilities are rapidly improving. Experts anticipate China will eventually develop its own comparable models, driven by the inevitable trend of increasing AI capabilities in the cybersecurity sector. This competition highlights a global race to enhance AI-driven cyber defense.