Art.13

Article 13 — Transparency and Provision of Information to Deployers

High-Risk Systems SME Relevant ~2 min read · 473 words

Article 13 requires providers of high-risk AI systems to ensure their systems are transparent enough for deployers to understand and use them correctly. This means providing clear, accessible instructions for use — covering what the system does, what it cannot do, and how to interpret its outputs.

! High compliance impact for SMEs

WHAT THE ARTICLE IS ABOUT

Making high-risk AI systems understandable to those who deploy them

Article 13 establishes transparency obligations for providers of high-risk AI systems. It recognises that a deployer who does not understand what an AI system does — or what its limitations are — cannot use it responsibly. The article mandates that providers supply the information deployers need to make informed decisions and comply with their own obligations.

WHAT IT SAYS

Clear instructions, honest limitations, interpretable outputs

High-risk AI systems must be designed and developed in a way that enables deployers to understand how the system works sufficiently to use it correctly and as intended
Providers must supply instructions for use that include: the identity and contact details of the provider, the characteristics and capabilities of the system, any known limitations, the level of accuracy and the metrics used to measure it, and the expected lifetime of the system
Instructions must also cover the circumstances under which the system may fail or produce inaccurate outputs, and any human oversight measures that should be applied
Where the AI system processes personal data, the instructions must include information to assist deployers in meeting their data protection obligations
Instructions must be clear, complete, correct, and in a format accessible to deployers — including machine-readable format where appropriate

WHO IS AFFECTED

Providers owe transparency to deployers — deployers owe it to users

Providers of high-risk AI systems — must create and supply compliant instructions for use
Deployers of high-risk AI systems — must receive, read, and act on the information provided
End users who interact with AI-assisted decisions — benefit indirectly through the obligations this article places on deployers
Businesses evaluating AI vendors — Article 13 creates a standard of what information any reputable provider should be able to supply

WHAT IT MEANS FOR SMES

A checklist for what to demand from your AI vendor

If you are a deployer: Article 13 gives you the legal right to expect detailed, accurate documentation from your AI provider — if a vendor cannot supply this, that is a compliance red flag
Specifically ask vendors for: accuracy metrics, known failure modes, recommended oversight measures, and data protection guidance
If you are a provider: your instructions for use are a legal document, not a marketing brochure — they must be honest about limitations and failure conditions
The requirement to disclose accuracy levels is significant — if your system has known accuracy gaps in certain contexts, you must document and communicate this
Keep the instructions for use on file — deployers may need to show them to regulators or use them when conducting their own fundamental rights impact assessments under Article 27

Article 14 — Human oversight (instructions must specify what oversight is required)
Article 26 — Deployer obligations (deployers must use this information to comply)
Article 27 — Fundamental rights impact assessment (instructions for use feed into this)
Article 50 — Transparency to end users (separate but related transparency obligation)

← Back to the Guide