AI vendor risk is the risk an organisation faces when it buys a product that relies on AI models: how its data is used for training, which model providers process it, how outputs are controlled and how the vendor meets AI rules.
What AI vendor risk means
AI changes familiar vendor questions. Data may be used to train or improve models. New sub-processors appear when a vendor adds a model provider. Terms and model versions change more often than in traditional software. Outputs can be wrong in ways that are hard to detect.
The applicable rules are mostly familiar. GDPR Article 28 governs model providers that process personal data. The EU AI Act adds provider and deployer duties: transparency obligations have applied since 2 August 2026, and high-risk duties follow from 2 December 2027 or 2 August 2028. ISO/IEC 42001 and the NIST AI RMF give management frameworks.
Most of the evidence sits in documents you already collect, especially the DPA, the sub-processor list and the product terms, which were often signed before the AI features shipped. Read them again when the vendor announces AI features, because the processing they describe may no longer match the product.
What to check in vendor documents
- Whether customer data is used to train or improve models, and the contractual opt-out.
- Model providers listed as sub-processors, with their locations.
- Retention of prompts and outputs, including by the model provider.
- Notice of changes to models or model providers.
- The vendor's role under the AI Act and its approach to Article 50.
- Human oversight, and how errors in outputs are handled.
In ProcureCortex
ProcureCortex reads an AI vendor's DPA, terms and policies like any other vendor's documents, so a gap such as an unlisted model provider becomes a finding that cites the clause and page.
How an assessment worksRelated terms
- Model provider sub-processorA model provider sub-processor is a company whose AI model a vendor uses to deliver its service and which processes your personal data in doing so. Under GDPR it is a sub-processor that needs authorisation and flow-down terms.
- EU AI ActThe EU AI Act is Regulation (EU) 2024/1689, the EU's risk-based law on artificial intelligence. It bans some practices, sets duties for high-risk systems, general-purpose AI models and transparency, and splits obligations between providers and deployers.
- General-purpose AI model (GPAI model)A general-purpose AI model is, under Article 3(63) of the EU AI Act, an AI model trained on large amounts of data that displays significant generality, can competently perform a wide range of distinct tasks and can be integrated into many downstream systems.
- ISO/IEC 42001 (AIMS)ISO/IEC 42001 is the international standard that specifies requirements for an artificial intelligence management system (AIMS), covering how an organisation that develops, provides or uses AI governs it responsibly. Organisations can be certified against it.
- NIST AI Risk Management Framework (AI RMF)The NIST AI Risk Management Framework (AI RMF) is a voluntary framework from the US National Institute of Standards and Technology for managing the risks of AI systems, organised into four functions: Govern, Map, Measure and Manage.
Sources
- Regulation (EU) 2024/1689 (AI Act)eur-lex.europa.eu/eli/reg/2024/1689/oj
- Regulation (EU) 2016/679 (GDPR)eur-lex.europa.eu/eli/reg/2016/679/oj
- NIST, AI Risk Management Frameworkwww.nist.gov/itl/ai-risk-management-framework
Summarised for practitioners, not legal advice. Check the source text for your own obligations.