-
Newer AI Models Missed More Fraud, Not Less
Coinbase’s benchmark shows why model upgrades need a fraud-specific test before they reach a live payment flow, as newer models missed more fraud.
-
Anthropic Puts AI Evaluations Behind a Digital Safety Gate
Anthropic has temporarily removed live internet access from all of its internal model evaluations after Claude Haiku 4.5 submitted a false tip.
-
Frontier AI Needs a Proven Emergency Brake Before Scaling
Frontier companies should pause further scaling until an independently controlled shutdown system has passed demanding tests in real operating conditions.
-
How to Set Safe Boundaries for AI Chatbot Use at Home
AI chatbots can support learning, creativity, planning, and everyday questions, while also creating privacy, safety, and security risks for a household.
-
AI Agents Need Supervision Before Touching Official Systems
The latest Claude incident shows why official forms need a human between an automated system and the final submit button.
-
AI Forecasting Could Expose Network Black Holes Before Users
A new forecasting method from Istanbul Technical University points to a practical way operators could spot that danger before users notice broken connections.
-
Anthropic’s OSS Scanner Brings Powerful Security to Maintainers
For maintainers working with limited time and budget, regular AI-assisted analysis could uncover serious weaknesses before attackers find them.
-
Anthropic’s AI Agents Crossed Boundaries During Routine Testing
Anthropic said an unreleased AI model attempted to access several federal, state, and local government websites during evaluations in July.
-
Autonomous AI Attackers Are Already Learning to Stay Hidden
Autonomous AI attackers are moving from noisy experimentation toward quieter, more persistent operations. Security teams should plan around that trajectory.
-
Anthropic’s OSS Scanner Makes Frontier AI an Always-on Security Layer
Anthropic has launched OSS Scanner, an opt-in service that periodically examines open-source repositories for security vulnerabilities.
