The Malicious Use of Artificial Intelligence report

An arXiv-updated report by Miles Brundage and coauthors surveys threats from malicious uses of AI across digital, physical, and political domains, and recommends forecasting, prevention, and mitigation measures.

• Anthropic co-founder: AI 'kill switch' may need to be mandatory2026-09-14
• Security experts urge basics before external AI audits2026-09-17
• Zuckerberg distances Meta from AI slowdown calls2026-09-16
• Anthropic and Microsoft debate AI limits2026-09-16
• Trump comments on AI2026-09-16