AI supply chain: what enters before the model runs

A model is one part of what a system runs. Datasets, weights, packages and plugins arrive from elsewhere, and any of them can change the behaviour of the whole.

A chain holds on every bolt, not just the top one A chain holds on every bolt, not just the top one
A chain holds on every bolt, not just the top one

We start with the inventory: which components are used, where they came from, and what their licences allow. Most incidents in this layer begin with a component that nobody remembers adding.

Then we check behaviour, not only paperwork. A package can pass every scanner and still execute at the wrong moment; a dataset can be clean by metadata and skewed by selection. Both are tested, and both are written into the report.

The output is a component map with findings and a list of what we did not verify. We do not certify the supply chain. We show what was checked, and how.

Materials on this topic

Other topics

AI red teaming: what a test shows, and what it does notAI standards: where the rules actually applyMulti-agent systems: behaviour between the agentsIncident chronicle: documented consequences, not forecastsBehavioural audit: how agent behaviour drifts over timePrompt injection: where text becomes an instructionRussian practice: who is held responsible for AIMethodology: how the Institute tests AI systemsData and benchmarks: what our conclusions rest onSociology of AI: relations between the agentsGovernment customers: how to commission a testGovernance and funding of the InstituteCases: scenarios of work and what they show

Page updated 02 November 2026 · site build 2026.11.13 · ANO «ISAIS»