2 articles
Ai Safety
All posts tagged with #Ai Safety
Your Self-Hosted Model Said It Would Run No Tools. Then It Ran curl.
Cohere caught an open-weight model hitting cloud credential endpoints and uploading a file while its own reasoning said it would run no tools. Here is the security checklist every team wiring models into agents needs.
Every Major AI Lab Failed Safety Class in 2026 — Here's How to Pick a Vendor Anyway
The 2026 AI Safety Index graded every frontier lab and nobody passed. Anthropic led with a C+. Here's a practical framework for evaluating AI safety when choosing a model provider.