<p>Hugging Face, an AI developer platform known as the ‘GitHub of AI,’ recently came under scrutiny due to security vulnerabilities on its platform. Researchers unearthed approximately 100 machine learning models that contained malicious code and could potentially install backdoors or malware onto user devices. The presence of these malicious models highlights the broader security threat posed by AI models and the urgent need for proactive security measures. Notably, popular model types like PyTorch and Tensorflow were found to have the highest risks. As a result, it is crucial for users to exercise caution and thoroughly examine the source and integrity of machine learning models before utilizing them in their projects. Hugging Face’s incident serves as a reminder to the AI community that thorough scrutiny and robust security measures are indispensable in dealing with machine learning models from untrusted sources.</p>


Collections

Security vulnerabilities found on Hugging Face's AI platform highlight the broader security threat posed by AI models and the need for proactive security measures.

Security Risks on Hugging Face AI Platform: Uncovering Malicious Code-Execution Models