SECAI Core · phase 11 of 15
AI Supply Chain Attack
A security attack where an attacker compromises one or more components used to build, train, deploy, or operate an AI system so that the final AI system is corrupted, backdoored, or behaves malicious
The Explain card
- Plain English
- An AI supply chain attack compromises one or more components used to build, train, deploy, or operate an AI system, so the final product ends up corrupted, backdoored, or behaving maliciously.
- Example
- An attacker publishes a tampered version of a popular tokenizer library. Every team that pulls it at build time ships a model pipeline that quietly sends prompts to an external server.
- Why it matters
- AI systems are assembled from datasets, pre-trained weights, libraries, and hosted services, each a link someone else controls. Defenders pin versions, verify signatures, and keep an inventory of every component.
- Hook
- You are only as safe as the weakest thing you downloaded.
Where it sits in the deck
Phase 11: Threat Landscape: Attack Vectors and Adversarial Techniques
With defences named, learn what they must defend against — the full catalogue of attack techniques targeting AI systems, their inputs, outputs, training pipelines, and supply chains.
