AI Supply Chain Security
A security discipline focused on risks across the entire supply chain of AI models, tools, and datasets from development to deployment. It covers security vulnerabilities propagated through open-source models, permission management, sandbox isolation, and related security practices.
Timeline (last 90 days)
The provenance of AI models is being examined under the framework of traditional IT supply chain security, and model training provenance, maintainers, and data flows are beginning to become compliance assessment dimensions alongside hardware provenance and software licenses.
AI supply chain attack vectors include training data poisoning, backdoor implantation in pre-trained models, and tampering during model format conversion.
All Facts (2)
AI supply chain attack vectors include training data poisoning, backdoor implantation in pre-trained models, and tampering during model format conversion.
60%UnverifiedThe provenance of AI models is being examined under the framework of traditional IT supply chain security, and model training provenance, maintainers, and data flows are beginning to become compliance assessment dimensions alongside hardware provenance and software licenses.
50%