🇰🇵 North Korean Hackers Start Using AI-generated Images in Attacks

The Kimsuky group, linked to the North Korean government, used an AI-generated military ID card to trick South Korean officials into clicking on a phishing link.

🕵🏻 How the Scam Worked

AI tools typically refuse to generate fake documents—but security researchers have found that the filter is easy to bypass. The hackers simply asked ChatGPT to create a "design sample" instead of a "copy of an ID." The chatbot then produced a template that looked almost identical to a real Ministry of Defense pass ⤴️

Victims received emails from addresses that closely resembled official MoD accounts. They were asked to review the "design" of the military ID. To give feedback, they had to click a link, which secretly installed spyware on their computers.

⚠️ AI In The Wrong Hands

This isn't the first time North Korean hackers have used AI. For example, they previously used Claude to generate fake CVs, prepare for job interviews, and even complete work tasks—all to land remote tech jobs.

Once inside the company systems, the attackers stole sensitive company data. Anthropic blocked the suspicious accounts, but it appears that the tug-of-war between defenders and attackers is only heating up.

#cybersecurity #ChatGPT @hiaimediaen