Hackers Use ChatGPT-like AI for Cyberattacks: Kimsuky Group

Hackers are reportedly building AI tools similar to ChatGPT to automate cyberattacks and make them more convincing. A North Korea-linked group, Kimsuky, is setting up AI software on its own systems. South Korean cybersecurity company Genians found signs of this activity. This technology could help hackers analyze stolen information, automate tasks, and create more believable scams. Kimsuky is reportedly using tools like Ollama, GPT4All, and Msty, which run AI models locally. This allows them to avoid cloud-based AI services. Researchers also found signs of retrieval augmented generation technology, which lets AI systems search large document collections. AI can make fraudulent messages sound more natural and professional. Kimsuky may be exploring wider uses, including AI-assisted coding, speech-to-text features, and AI agent development. This could help automate parts of cyberattacks and process stolen information faster. Finance and cryptocurrency documents appearing to be AI-created were also found, designed to look like genuine reports. These could trick victims into opening malicious files or falling for scams. The bottom line is users should be very careful about unexpected messages, files, and financial offers online.