DATA SCIENTISTS TARGETED BY MALICIOUS HUGGING FACE ML MODELS WITH SILENT BACKDOOR

Summary from the TLDR InfoSec newsletter

Researchers discovered a malicious AI model on Hugging Face that could enable attackers to gain full control of users' machines. The model executes code after loading a pickle file, leaving victims unaware that their systems were compromised.