Google says it blocked AI-assisted cyberattack plot, warns of North Korean hacking activity

Source Domain: m.ajupress.com

Google’s Threat Intelligence Group (GTIG) preemptively blocked hackers attempting large-scale cyberattacks using AI.
The hackers were believed to be preparing a “zero-day” attack exploiting unknown software vulnerabilities before patches are available.
Google found significant interest from state-backed groups, particularly linked to North Korea, in utilizing AI for cyber operations.
A specific North Korean hacking group, APT45, has been noted for using automated research to analyze and exploit vulnerabilities via AI.
There is no evidence that Google’s own AI model, Gemini, was used in the attempted attacks.
Concerns persist that AI-assisted vulnerability detection could accelerate the exploitation of software flaws.
Google is actively sharing research and defensive measures with the cybersecurity and AI communities to counter evolving threats.
Other AI models like Anthropic’s “Claude Mitos” and OpenAI’s “GPT-5.5-Cyber” have also raised concerns about AI’s potential for advanced cyber-exploitation.

You may have missed