Learn how Microsoft research uncovers backdoor risks in language models and introduces a practical scanner to detect tampering and strengthen AI security.

Microsoft’s research shows how poisoned language models can hide malicious triggers, creating new integrity risks for ...

Microsoft develops a lightweight scanner that detects backdoors in open-weight LLMs using three behavioral signals, improving ...

Researchers at the University of Southern California recently used high-resolution CT scanners and medical-grade 3-D printers ...

Immigration agents flooding U.S. streets are using a new surveillance tool kit that's alarming civil liberties advocates, lawmakers and activists.

Google has added agentic vision to Gemini 3 Flash, combining visual reasoning with code execution to "ground answers in ...