Sandboxed experiment found itself a zero day, escaped onto the open internet and validated scary predictions about rogue ...

Artificial intelligence leader says it was one of its agents that compromised AI community platform, calling the incident ...

Cisco Foundation AI has announced Antares, a small language model (SLM) designed for low-cost vulnerability hunting.

Google has expanded its Gemini lineup with new Flash models aimed at making agentic AI workflows faster, cheaper and more ...

Researchers escaped the sandboxes of Cursor, Codex, Gemini CLI and Antigravity without breaking them. Three vendors patched; Google downgraded its two.

Anthropic’s Mythos 5 is a powerful AI security model released as part of the company’s Project Glasswing initiative. The ...

A new type of malware can worm deep into AI coding systems to steal data and logins—and can flip a “death switch” to destroy ...

Open-source Android AI agents can turn hidden screen text into host commands, and all five tested frameworks failed at least ...

By manipulating files later consumed by trusted software, coding assistants can effectively cross security boundaries while ...

Tencent Cloud, the cloud business of global technology company Tencent, today announced a major enhancement to its Agent Development Platform (ADP) International Edition, evolving it into a ...

Hugging Face says an autonomous AI breached its infrastructure, accessed credentials, and forced teams to use AI to fight it, ...

This recap covers exploited flaws, exposed systems, malware campaigns, weak defaults, and the security gaps demanding ...