OpenAI builds secure sandbox for Codex on Windows
OpenAI launches secure sandbox for Codex on Windows, enabling safe coding with controlled file/network access. Targets US developers.
-
-
Microsoft AI System Discovers 16 Windows Flaws in Patch Tuesday
Microsoft’s AI tool MDASH identified 16 Windows vulnerabilities fixed in May 2025 Patch Tuesday — a sign of AI’s growing role in proactive cybersecurity defense across millions of devices.
-
OpenAI Enters Cybersecurity Market With Daybreak
OpenAI launches Daybreak, entering the enterprise cybersecurity market — putting the AI lab in direct competition with CrowdStrike, Palo Alto Networks, and AI-native security startups.
-
New Benchmarks Expose Security Gaps in Enterprise AI Agent Access Controls
New research exposes a dangerous AI agent failure mode: when access controls silently filter restricted data, agents give confident but materially incomplete answers. A new 72-task benchmark measures the risk.
-
White House ‘Studying’ AI Security Executive Order
The White House is studying an executive order on AI security that could specifically target Anthropic and reshape federal AI procurement and national security protocols.
-
Anthropic Launches Enterprise Security Tool Ahead of Mythos Model Release
Anthropic enters enterprise cybersecurity market with new security tool ahead of its Mythos model release, joining Microsoft, Google & startups in the fast-growing AI security space.
