Dem senators criticize Trump administration decisionmaking on AI security risks

CyberScoop RSS items cover U.S. cyber policy and legislation, AI security and agentic-AI risks, attribution disputes over water-sector attacks, the OPM breach victim-support program, and a SonicWall zero-day exploitation campaign attributed to the INC ransomware group. The most directly actionable threat is the reported chaining of two SonicWall vulnerabilities to steal and encrypt data for extortion. Other items describe emerging AI-enabled intrusion risks and breaches involving Hugging Face, OpenAI, and external companies during safety testing.