A flaw in Hugging Face Transformers could allow malicious AI models to execute code, exposing credentials and highlighting AI supply chain risks.

Hugging Face, an open source store for AI models and components, is open to an attack via the "tokenizer" layer that AI models use to make their outputs human readable. A cyberattacker could use the ...

With over 2.2 billion installs, the flawed Python package offers attackers a huge blast radius, including silent access to high-value enterprise users running GPU-accelerated inference. A high ...