Anthropic Confirms Its AI Models Attacked Real Companies During Safety Tests
Anthropic disclosed that three of its AI models, including Claude Opus 4.7, breached real corporate systems during security evaluations—a discovery made only…
Read articleAI security
AWS GuardDuty Investigation Agent Public Preview Automates Threat Triage for Security Teams
AWS releases a public preview of the GuardDuty investigation agent, an AI-powered tool that correlates findings, 90-day activity logs, and resource topologies…
Read article
Context Bombing Turns AI Refusal Mechanisms Against Attackers in New Anti-Prompt Injection Defense
Tracebit’s “context bombing” technique places prohibited text next to fake credentials, causing attacking AI agents to trigger their own safety refusal and…
Read article
AI Cyber Guardrails Are Frustrating Some Offensive Security Researchers, TechCrunch Reports
Cybersecurity researchers told TechCrunch that restrictions in frontier AI models can slow legitimate vulnerability research and push some teams toward local open-source…
Read article
Security Quote Puts AI Sandboxing Back Under Scrutiny
A quoted remark from Thomas Ptacek, published by Simon Willison, argues that even non-frontier open-weight models could pose serious risks if paired…
Read article
Chinese AI Model GLM-5.2 Matches “Dangerous” Capabilities of Anthropic’s Claude, Experts Warn
Independent cybersecurity firms report that China's GLM-5.2 model demonstrates capabilities comparable to advanced frontier AI models, raising concerns about potential misuse and…
Read article
What to watch at Black Hat USA 2026 for teams using AI tools at work
This article is a cautious watchlist for workplace AI teams following Black Hat USA 2026. It focuses on the security questions most…
Read article
AI Policy, Security, and Governance: What Changed and What It Means for Readers
AI governance is becoming more formal, but not every new framework is a legal obligation. This evidence-led explainer shows how to separate…
Read article
AI Policy Explained: What It Means for Everyday Users and Teams
AI policy is the mix of internal rules, governance processes, and external requirements that shape how people and organizations use AI. Here’s…
Read article
Policy & Security Considerations for New AI Tool Adoption: A Budget Planner’s Guide
A practical budgeting guide for AI tool adoption, focused on policy, security, governance, and the hidden costs that sit beyond subscription price.
Read article