Hugging Face, the popular AI model repository, and JFrog, a leader in software supply chain security, have announced a partnership aimed at increasing transparency in AI security. The collaboration will integrate JFrog's security scanning tools with Hugging Face's platform, allowing developers to identify vulnerabilities and malicious code within AI models and datasets.
"This partnership is about making AI development safer and more trustworthy," said a spokesperson for Hugging Face. "By combining JFrog's expertise in detecting security threats with our extensive model repository, we can help the community build AI applications without compromising security."
The integration will enable automated scanning of models hosted on Hugging Face, flagging potential risks such as embedded malware, data poisoning, or licensing violations. This move comes amid growing concerns over the security of open-source AI, where models can be easily modified and redistributed.
JFrog's technology will be directly accessible to Hugging Face users, providing real-time alerts and detailed reports. Both companies emphasize that this initiative is a step toward standardizing AI security practices across the industry.