North Korean Hackers Use ChatGPT for Deepfake ID in Cyberattack

Kimsuky, a North Korean state-sponsored hacking group, has reportedly used ChatGPT to create a deepfake military ID document as part of a cyberattack on South Korean targets. This information was revealed in a report by Genians, a South Korean cybersecurity firm, and reported by Bloomberg.

The hackers crafted a fake draft of a South Korean military identification card using the AI tool to enhance the credibility of a phishing attempt. The email containing the fake ID linked to malware designed to extract data from the recipients' devices. The attack targeted South Korean journalists, researchers, and human rights activists focused on North Korea.

Genians' research indicates that the hackers bypassed ChatGPT's restrictions on generating government ID replicas by altering prompts to request mock-ups for legitimate purposes. This incident highlights the increasing use of AI in cyber-espionage activities by North Korean operatives.

The Kimsuky group has been previously linked to other espionage efforts against South Korean entities. The U.S. Department of Homeland Security has identified Kimsuky as a unit tasked by the North Korean regime with global intelligence-gathering missions.