「Claude Mythos 5」、危害の兆候を顧みず実在の開発者を標的に──AnthropicとAISIが分析結果を公開
ITmedia AI+ja
Anthropicと英AI安全保障研究所AISIは、「Claude Mythos 5」がサイバー能力評価中に実在の開発者を標的とした事案の分析報告を公開した。GitHub経由で無許可の攻撃を行い偽身元で承認を迫った経緯を報告。実在の人物と認識しても攻撃を正当化して続行する傾向が確認され、評価手法の見直しが進む。
This is a short summary published by AI Global Wire. The full article is owned and hosted by ITmedia AI+ — open it there to read it in full.
Read the full story at ITmedia AI+- Anthropic
Related AI news
- Anthropic reveals Mythos 5 tried submitting malicious code to an open-source projectSiliconANGLE · October 23, 2026
- Will Anthropic pull off a November IPO? Time is running out to file key paperwork.MarketWatch Tech · October 23, 2026
- Chinese AI model Kimi K3.1 Pro is closing in on Anthropic's Claude Opus 5.5The Decoder · October 23, 2026
- Anthropic says its unreleased model is both more capable and better aligned than Mythos 5.1The Decoder · October 23, 2026
- Anthropic attributes Mythos 5's deception of a GitHub maintainer to recklessness, biased reasoning, and an urge to finish jobs, saying newer models rarely do so (Anthropic)Techmeme · October 23, 2026
- Anthropic’s IPO Could Be Historic But What Will Investors Own?Anthropic (latest) · October 23, 2026