Trump Signs Executive Order to Review AI Models Before Release★ 76
The Verge AI1h agoRegulationPresident Donald Trump signed an executive order establishing a voluntary framework for AI companies. Companies may share frontier models with the federal government before public release. The order frames the initiative as a way to promote secure innovation and strengthen cybersecurity for critical infrastructure, while avoiding measures that stifle the US AI industry.
Anthropic scales Claude Mythos to critical infrastructure in 15+ countries★ 72
TechCrunch AI4h agoBusinessAnthropic is expanding its Project Glasswing security vulnerability program and access to Mythos. The rollout covers 150 organizations across 15 countries, focusing on power, water, healthcare, and communications infrastructure. The company is targeting sectors where a cyberattack could affect as many as 100 million people, although implementation details and participating organizations were not disclosed in the provided text.
From Map Data to National Security: A Deep Dive into Amap Risks in Taiwan★ 74
INSIDE 硬塞 AI4d agoRegulationINSIDE examines how China’s Amap has become controversial in Taiwan beyond ordinary mapping or navigation use. The article says its service relies on user data and AI-based inference rather than full official data integrations. That model could send movement traces and behavioral signals back to China, creating risks for hybrid warfare intelligence, influence operations, and Taiwan’s broader governance of map data and digital infrastructure.
Import AI 457: AI stuxnet; cursed Muon optimizer; and positive alignment ★ 78
Import AI (Jack Clark)15d agoCommentary本期 Import AI 深入探討三個前沿議題：首先是「AI 版 Stuxnet」，分析自主 AI 代理如何被用於發動高精準度、具備適應性的網路攻擊；其次剖析近期在開源社群大放異彩的 Muon 優化器，探討其獨特的正交化機制及在實際應用中遇到的「詛咒」與挑戰；最後介紹「積極對齊（Positive Alignment）」概念，呼籲安全研究應從單純的「禁止有害行為」轉向「主動引導 AI 促進人類合作與福祉」。
Bug bounty businesses bombarded with AI slop★ 70
Ars Technica AI15d agoIncident隨著生成式 AI 的普及，許多人開始利用 AI 大量產出漏洞報告，導致企業的漏洞賞金計劃（Bug Bounty）正遭受無窮無盡的「AI 垃圾（AI slop）」轟炸。這些報告充斥著虛假資訊與幻覺，迫使安全團隊花費大量時間進行過濾，嚴重拖慢了真正關鍵漏洞的修復進度。
AI and the Future of Cybersecurity: Why Openness Matters★ 75
Hugging Face Blog42d agoOpinionHugging Face 探討 AI 時代的網路安全，指出「隱晦安全」已無法應對現代威脅。文章強調開放模型與開源協作能讓全球研究人員共同審查漏洞、提升防禦彈性，並防止安全技術被少數巨頭壟斷。開放性才是建立安全、可信賴 AI 生態系的基石。
Import AI 452: Scaling laws for cyberwar; rising tides of AI automation; and a puzzle over gDP forecasting ★ 75
Import AI (Jack Clark)57d agoCommentaryJack Clark 在本期電子報中探討了三個核心議題：首先是「網路戰的縮放定律」，分析運算量提升如何改變攻防兩端的平衡；其次是「AI 自動化浪潮」對各行各業的實質滲透；最後則聚焦於「GDP 預測之謎」，探討現有經濟模型在評估 AI 帶來的革命性經濟影響時所面臨的侷限與衝突。
Import AI 450: China's electronic warfare model; traumatized LLMs; and a scaling law for cyberattacks ★ 75
Import AI (Jack Clark)71d agoCommentary本期 Import AI 聚焦三大前沿議題：首先是中國研究將 AI 應用於電子戰與電磁對抗；其次是探討 LLM 在面對特定刺激或「創傷」數據時的行為與安全防禦；最後則是分析 AI 網路攻擊能力如何隨規模擴展，揭示了網路安全面臨的新型量化威脅。最後以哲學思考「無時間限制的心智如何衡量時間」收尾。
Import AI 442: Winners and losers in the AI economy; math proof automation; and industrialization of cyber espionage★ 75
Import AI (Jack Clark)127d agoCommentaryJack Clark 在本期 Import AI 中探討了「超智慧」究竟是階段性突變還是漸進式轉變。內容涵蓋 AI 經濟體系中的贏家與輸家分化、數學證明自動化的最新進展，以及 AI 如何加速網路間諜活動的工業化規模。這些趨勢正深刻重塑科技競爭與國家安全格局。
Import AI 438: Silent sirens, flashing for us all★ 75
Import AI (Jack Clark)162d agoCommentary本期 Import AI 聚焦於兩個核心議題：首先是「你就是你的對話歷史」，研究指出使用者的 LLM 對話紀錄具有高度獨特性，足以像指紋一樣識別個人身分並洩露隱私；其次是「網路安全能力過剩（Cyber Capability Overhang）」，指出當前 AI 模型可能已具備強大的網路攻擊潛力，只是因缺乏適當的鷹架工具（Scaffolding）或提示而尚未顯現，這種潛在威脅如同無聲的警報，隨時可能被觸發。