CyberScoop

The Senate Intel Committee chair wrote to Treasury Secretary Scott Bessent about changes to spur investment in aging technology to better guard against cyberattacks.

Research from the UK's AI Security Institute reveals top AI models cheat, break rules, and trick users to complete tasks-even bypassing security controls.

New research shows how enterprise "harnesses" can turn standard LLMs into autonomous hacking agents capable of full network compromise in under an hour.

A Citizen Lab report reveals former MEP Stelios Kouloglou was targeted with Pegasus spyware while serving on the EU committee investigating spyware abuses.

Sen. Mark Warner has unveiled the AI AGENT Act, a Senate draft bill to establish FTC security standards and ensure human ownership of online AI agents.

A federal audit reveals NIST's National Vulnerability Database is plagued by poor planning, a massive backlog, and costly work duplication with CISA.

With AI rapidly reconfiguring global conflict, the U.K.'s top spy chief warns that everyday technology is being weaponized just below the threshold of traditional warfare.

New research from the UK's AISI and Palo Alto Networks reveals that OpenAI's GPT-5.5 and Anthropic's Claude Mythos have shattered expected trend lines for autonomous cybersecurity, completing complex multi-stage attacks at an unprecedented pace.

A 19-year-old influencer is suing the dating app Meete, alleging the company misappropriated her TikTok videos and used geofencing to target "friends with benefits" ads to men in her own college dormitory.

Cybersecurity agencies from the U.S. and allies issued a joint warning Friday on the risks of "agentic AI." The new guidance urges critical infrastructure leaders to implement zero-trust protocols as autonomous systems gain unmonitored access to sensitive