Taiwan Says It Was Targeted Last Month in AI-Driven Hacking Campaign
Source: Reuters, Ben Blanchard and Raphael Satter, 2026-08-13.
Reuters’ account adds two elements the Dream Security technical report does not carry: an official government confirmation, and an independent caution against over-reading the autonomy claim.
Government confirmation
Taiwan’s Ministry of Digital Affairs said its cybersecurity monitoring units detected an “abnormal attack” against government agencies in July 2026, issued warning alerts from 2026-07-20, and that the attack showed “clear characteristics of an overseas source,” combining manual operations with AI agent-assisted attacks, naming “Open Claw” specifically. The ministry said affected units “have successively completed their handling” and did not name China; China’s Taiwan Affairs Office did not respond to Reuters’ request for comment. The statement followed the Dream Security disclosure by one day — the Financial Times was first briefed and identified the target as Taiwanese, since Dream itself declined to name the government when approached by Reuters.
Caution against overclaiming autonomy
Cris Thomas, a security advocate at Semgrep, is quoted directly: “There’s still a human in there somewhere. Somebody had to choose who to attack, had to establish an objective and give it a directive. It’s not totally 100% autonomous. There was a capable operator in charge that did that.” Reuters frames this alongside the broader ramp-up in AI-assisted hacking following the release of frontier reconnaissance-and-exploitation-capable models, naming Anthropic’s Mythos as context for the trend — not as the tool this campaign used, which Dream’s report identifies as Hermes and OpenClaw.
Bearing on wiki positions
Corroborates Dream’s account from the victim side and supplies the human-operator caution that keeps the incident correctly framed as deliberately constructed and operator-directed, not fully autonomous — see the comparison in Offensive Agent Collective and GTG-1002.