Malware Alert: Cybercriminals Hide New Threat in AI & ML Models on PyPI!
ReversingLabs uncovers a sneaky new malware tactic targeting Alibaba AI Labs users by stashing malicious code inside AI/ML models on PyPI. These packages might look like a Python SDK but are actually infostealers in disguise. The takeaway? Don’t let your guard down around Pickle files or you might get yourself in a real pickle!
Hot Take:
Who knew that AI models could be a hacker’s best friend? With malware lurking in AI/ML models on PyPI, it seems cybercriminals have finally found a way to make artificial intelligence work for them. Hats off to ReversingLabs for catching these sneaky Pickle files before they could wreak more havoc than a cat on a keyboard!
Key Points:
- Cybercriminals are hiding malware in AI/ML models on the Python Package Index (PyPI).
- Three malicious packages targeted Alibaba AI Labs users by masquerading as Python SDKs.
- Packages dropped an infostealer hidden inside a PyTorch model, exploiting the Pickle file format.
- 1,600 downloads occurred before the packages were removed, likely aided by phishing or social engineering.
- The rise of AI/ML in software supply chains presents new opportunities for cyber attacks.
Already a member? Log in here