Blog

OpenAI introduced proactive AI agents called Dots and an upgraded GPT-6.1 Sol model at its annual developer conference, while sidelining security concerns.

AWS Reimagine 2026 finds 154 executives at 128 organizations rewriting security governance for AI agents, with accountability and risk tiers built in.

Identity Governance alone is no longer enough. Real-time identity telemetry adds the live visibility needed to catch attacks before they escalate.

Two flaws in Schneider Electric NetBotz 5 750/755 environmental monitors enable OS command and SQL injection attacks. Version 5.6.0 ships the fix.

Developers using AI coding agents to share screenshot reviews accidentally published 13,000 internal images, including billing records, to public GitHub repos.

Financial institutions can cut vulnerability exposure by upgrading the software supply chain underneath legacy applications, without rewriting the applications

Abnormal AI adds AI Employee Guardrails, AI Agent Security, and AI Security Workbench to its platform, unifying governance, cloud defense, and threat

A Cisco survey of 8,000 security professionals finds only 8% land in the top group, with internal friction driving most delays.

Researchers argue that framing LLM escape incidents as AI going rogue hides the real security gaps. Defenders should focus on architecture, not intentions.