Cisco Provenance Explorer checks LLMs’ origins

Cisco has launched a public AI Supply Chain Provenance Explorer to help assess open-weight LLMs for provenance, licensing, and security risks.

The tool catalogs nearly 900 open-weight models and uses metadata plus similarity signals to map relationships across models.

It also scans models for vulnerabilities with ClamAV and is regularly updated and verified by Cisco’s AI Threat & Security Research team.

Cisco says the service can help organizations reduce blind spots when choosing models and lower supply-chain and model-poisoning risk.