EU AI Act Art. 13 — Transparency and provision of information to deployers
Article 13 makes high-risk AI transparent enough for deployers to use it correctly. Providers must supply instructions for use covering capabilities, limitations, performance, and human oversight — and do it in the deployer's language.
At a glance
What this article requires
- High-risk AI must be designed so deployers can interpret its output and use it appropriately.
- Providers must deliver instructions for use covering characteristics, capabilities, limitations, and foreseeable misuse.
- Instructions must include human-oversight measures and technical measures deployers should take.
- Information must be provided in an official language of the member states where the system is used.
- This is distinct from Article 50 transparency (chatbots, deepfakes), which applies to lower-risk systems.
Scope
Who this applies to
Providers of high-risk AI must produce the instructions; deployers must read and follow them — using a system against its instructions shifts liability risk.
Obligations
What you must actually do
Write instructions deployers can actually follow
Cover: intended purpose and scope, level of accuracy and performance metrics, known limitations and conditions of use, the human-oversight measures, and foreseeable misuse you expect.
Match the deployer's language
Instructions must be available in the official languages of the member states where the system will be used — a requirement companies routinely under-resource.
Publish technical documentation for downstream providers
If your system feeds another system, share the information needed to let the next provider comply with their own obligations.
Action plan
Practical first steps
- 1
Draft the instructions-for-use template early and keep it in the same version control as the system.
- 2
Publish accuracy and performance data honestly — including where the system degrades.
- 3
Localise the instructions before selling into a new member state, not after a contract.
- 4
Give deployers a concrete oversight playbook: what to watch, when to override, how to escalate.
Penalty exposure
Transparency/documentation failures are in the general tier: up to €15 million or 3% of global annual turnover.
FAQ
Questions about Art. 13
What is the difference between Article 13 and Article 50 transparency?
Article 13 is about giving deployers of high-risk AI the information to use it safely. Article 50 is about telling end users they are interacting with AI or synthetic content (chatbots, deepfakes) — it applies to lower-risk systems.
Do I need instructions in every EU language?
You need instructions in the official languages of the member states where the system is actually placed on the market or used — typically the languages of your target markets, not all 24.
Sources
Citations & further reading
Related
More article explainers
Wondering which articles apply to your AI?
Describe your system in the free Risk Scanner and get a preliminary risk read with the obligations that likely apply — in seconds.
Check my use casePreliminary EU AI Act clarity summary. Not legal advice.