Interest has grown in ARTEX, an AI-based penetration testing tool developed in China, after it was reported to have been used in cyberattacks targeting South Korean financial institutions. ARTEX is an autonomous penetration testing system built on multi-agent large language model technology and had been offered as open source. Developer Autumn-27 has moved the tool to private access, saying the attacks were unrelated and that malicious use violates its intent, and said updates, new releases and maintenance support would stop.

AI can identify people behind anonymous or pseudonymous online accounts by analysing public posts and combining web searches with inference, researchers found. A paper published in February described an automated process of extraction, search, inference and calibration. In tests on 338 Hacker News users, the system correctly identified 226, with 67 percent recall and about 90 percent precision. Researchers said costs were about $1 to $4 per profile and under $2,000 overall, but results do not mean all anonymous accounts can be identified.