Article 26(4)
UpcomingConditional timingEnsure Controlled Input Data Is Relevant and Sufficiently Representative
Applies to Deployer; High-Risk AI, where deployer controls input data.
- Actors
- Deployer
- AI class
- High-Risk AI
- Themes
- Third Parties & Supply ChainGovernance & Accountability
Tracker Networks Guidance
Where your organization controls the input data used by the high-risk AI system, ensure those inputs are relevant and sufficiently representative for the system's intended purpose.
Official text
4. Without prejudice to paragraphs 1 and 2, to the extent the deployer exercises control over the input data, that deployer shall ensure that input data is relevant and sufficiently representative in view of the intended purpose of the high-risk AI system.
Timing depends on the system
- 2 Dec 2027 — Article 6(2) / Annex III high-risk AI
- 2 Aug 2028 — Article 6(1) / Annex I Section A high-risk AI
- 2 Dec 2027 — Pre-existing Annex III high-risk AI type/model first placed on the market or put into service before 2027-12-02
- 2 Aug 2028 — Pre-existing Article 6(1) / Annex I high-risk AI type/model first placed on the market or put into service before 2028-08-02
- 2 Aug 2030 — Pre-existing high-risk AI intended to be used by public authorities
Suggested controls
High-Risk AI Data Governance and Bias Management
Govern training, validation and testing data for suitability, quality, representativeness and bias management.
High-Risk AI Deployment Governance
Govern deployer use against provider instructions, input-data requirements, monitoring, human oversight and escalation.
Related risks
AI Data Quality, Representativeness or Bias Failure
Training, validation or testing data may be unsuitable, poor quality, unrepresentative or biased for the intended high-risk AI use.
Improper High-Risk AI Deployment
A deployer may operate high-risk AI contrary to provider instructions, without qualified oversight, suitable inputs or required monitoring.