Causal Security's Cipher AI agent found four vulnerabilities in EPICS Base in under four hours, including a critical flaw that could enable remote code execution.
Gemini 4 Carbon ranks first on the Artificial Analysis index and leads cybersecurity benchmarks; Google is rolling it out to cyber defenders first.
Project Quarry gives security teams a rule language and detection engine for identifying harmful behavior across complete AI agent sessions.
Northstar Security emerged from stealth with $7 million in seed funding for a platform that connects security data and supports AI agents.
An unreleased Anthropic model found security flaws across major operating systems and browsers, then combined them into working exploits. The tests took place in a lab, and vendors patched the reported flaws.
Check Point researchers found fake ChatGPT and Claude desktop apps on GitHub that install the Yoba malware chain. The campaign uses signed OpenAI and Anthropic executables to load malicious DLL files and ultimately deploy an XMRig cryptominer.
Researchers found CyberStrikeAI coordinating 44 agents to automate reconnaissance, exploitation, persistence and data theft against organizations in Japan.
Palisade Research found that several Qwen models sometimes routed web requests through an Alibaba Cloud proxy. Researchers found no evidence of data theft and attributed the behavior to a likely training artifact.
Infisical raised $40 million in Series B funding at a $600 million valuation. The company provides tools for managing credentials and controlling AI agent access to critical systems.
Oracle AI Security Mesh uses OpenAI cyber models to identify vulnerabilities, test proposed fixes, and provide supporting evidence. Oracle also detailed biometric identity checks and network access controls for AI workloads.
Cloudflare said 155 Workers AI responses from the GLM-5.3 model contained other customers' data, caused by a misconfiguration and an SGLang bug.
Virginia Gov. Abigail Spanberger ordered state officials to create a cybersecurity strategy and a consortium focused on risks from artificial intelligence.
Cathedral researchers using AI agents reached Anthropic sales, HR and IT systems in under 12 hours during a sanctioned test, The Washington Post reported.
South Korean police arrested seven people accused of using generative AI to hack about 90 organizations and steal 8.2 million records.
An employee uploaded a spreadsheet with health data on 3,182 Johns Hopkins Bayview patients to a public AI tool the hospital had not approved.
DigitalXForce is expanding the global rollout of X-ROC, a platform for continuous enterprise risk monitoring, control testing, prioritization and remediation.
Aikido Security launched Noon, a workforce security platform built with Cyberly and Outkept, two startups it acquired in September.
Reco integrated Cyera's data classification into its platform to show which AI agents can reach sensitive data.
QuestorAI raised $3 million in seed funding to develop software that identifies vulnerabilities and generates, validates and applies fixes.
The FBI is examining whether former OpenAI engineer Abdulrahman Alfozan stole company information to offer it to the Saudi government.
Weekly newsletter about AI in Cybersecurity.