Today, Hugging Face adds a new model on average every 7 seconds, and the platform now hosts nearly 1.9 million models available to developers worldwide. This unprecedented scale — driven by contributors globally, spanning both trusted institutions and independent creators — fuels a wave of innovation while also reinforcing the need to secure the AI supply chain.
As highlighted in our previous analysis, AI supply chain risks now permeate every stage of the AI lifecycle — from vulnerable software dependencies and malicious or backdoored model files to poisoned or non-compliant datasets. Given this complexity, it is increasingly challenging for any single organization to address these issues alone. Effective security of the AI landscape requires close collaboration across the community to secure AI.
Elevating AI Supply Chain Security with Hugging Face
At Cisco, we are on a mission to help every organization on the planet securely execute their AI strategy. Today, we’re taking this mission a step further. We’re excited to announce a strategic relationship between the Foundation AI team at Cisco and Hugging Face, bringing together the world’s leading AI model hub with Cisco’s expertise in securing digital infrastructure.
As part of this expanded collaboration, Cisco Foundation AI will provide the platform and scanning of every public file uploaded to Hugging Face — AI model files and other files alike — in a unified malware scanning capability powered by custom-fit detection capabilities in an updated ClamAV engine.
By combining Hugging Face’s central role in open-source AI with Cisco’s comprehensive malware scanning capabilities, this enables more rigorous model vetting, early detection of vulnerabilities, and shared threat intelligence — building greater trust and stronger security across the entire AI ecosystem.
“We are thrilled to partner with Cisco Foundation AI to help secure Hugging Face users. We have been scanning files with ClamAV, the free and open source malware detection scanner from Cisco Talos, for a few years. With ClamAV’s new update we can now provide comprehensive protection against both traditional malware and threats unique to AI models—all with a single tool. We are grateful to Cisco to becoming our partner to scan all files uploaded to Hugging Face. By combining our leadership in open-source AI with Cisco’s deep cybersecurity expertise, we’re empowering organizations and individuals worldwide to adopt AI with confidence”
Julien Chaumond, CTO, Hugging Face
In addition, as a result of our collaboration, we are democratizing AI model antimalware:
- ClamAV can now detect malicious code in AI models– We are releasing this capability to the world. For free. In addition to its coverage of traditional malware, ClamAV can now detect deserialization risks in common model file formats such as .pt and .pkl (in milliseconds, not minutes). This enhanced functionality is available today for everyone using ClamAV.
- ClamAV is the only antivirus engine focused on AI risk in VirusTotal– ClamAV is the only antivirus engine to detect malicious models in both Hugging Face and VirusTotal – a popular threat intelligence platform that will scan uploaded models.
We’re proud to deliver our work on AI supply chain security to Cisco customers and now, the greater AI and security community. More is on the way to help protect AI developers from supply chain risks.
Learn More
The Cisco Foundation AI team recently launched Cerberus, a 24/7 guard for the AI supply chain. Cerberus inspects models as they enter Hugging Face, sharing results in standardized threat feeds that Cisco Security products use to build and enforce granular access policies for the AI supply chain.
With the release of ClamAV 1.5, Cisco brings deeper visibility into the AI model supply chain to the security community. ClamAV 1.5 adds native support for identifying AI model files during scanning to allow for model-specific detection logic and safer handling of embedded threats. Together with our signature updates (which do not require ClamAV 1.5) to ClamAV, ClamAV is now positioned as a foundational tool for securing the emerging AI model ecosystem. These capabilities are also available across the Cisco portfolio of products with our Talos threat intelligence services.
Users of Cisco Secure Access can configure how to provide access to Hugging Face repositories, block access to potential threats in AI models, block AI models with risky licenses, and enforce compliance policies on AI models that originate from sensitive organizations or politically sensitive regions.
We previously launched protections for Secure Endpoint, Secure Email Threat Defense, Secure Access and Secure Firewall. All existing users of Cisco Secure Endpoint and Email Threat Defense are protected against malicious AI Supply Chain artifacts.
For more information on the Foundation AI team, check out our website and feel free to send us a message!
We’d love to hear what you think! Ask a question and stay connected with Cisco Security on social media.
Cisco Security Social Media
Share:
