AI Tools Help North Korean Hackers Steal Millions

State-sponsored hackers from North Korea are increasingly turning to artificial intelligence to supercharge their cybercrime operations, according to a new report from cybersecurity researchers. One group, believed to be linked to the Lazarus collective, used AI tools across nearly every stage of their attack chain, from writing malware code to creating convincing fake company websites. The result: approximately $12 million stolen in just three months. The report details how the hackers employed AI for 'vibe coding'—a term describing the use of AI to generate functional code based on loose descriptions rather than precise instructions. This allowed them to rapidly develop custom malware variants that could evade traditional antivirus detection. More creatively, the group used AI image generators to create realistic profile photos for fake executives and AI text models to write professional-sounding business emails, making their social engineering campaigns far more convincing. Perhaps the most sophisticated tactic involved AI-generated company websites. The hackers would identify legitimate businesses, then use AI to clone their web presence with subtle modifications—such as changed bank account details on invoice pages. These fake sites were then promoted through targeted phishing campaigns, tricking suppliers and customers into making payments to the hackers' accounts. The entire operation was highly automated, with AI managing everything from initial reconnaissance to final money laundering instructions. This case represents a worrying escalation in cybercrime capabilities. While state-sponsored hacking groups have always been resourceful, AI tools dramatically lower the barrier to entry for complex attacks. The report's authors warn that similar tactics are likely being adopted by other nation-state actors and even criminal gangs. For businesses, the recommendation is clear: implement zero-trust security models, verify all payment requests through independent channels, and invest in AI-powered defense systems that can detect the subtle patterns of AI-generated attacks. The era of AI-versus-AI cybersecurity has truly begun.