These tools and metrics are designed to help AI actors develop and use trustworthy AI systems and applications that respect human rights and are fair, transparent, explainable, robust, secure and safe.
EU AI Regulation Decoded — EU AI Act obligations and evidence reference
The EU AI Act Obligation-to-Evidence Dataset is an open, machine-readable dataset that maps obligations under the EU AI Act to the specific evidence an auditor would expect to see. It is published by EU AI Regulation Decoded, an independent practitioner project. Most EU AI Act resources explain what the law says. This dataset addresses a more practical question: which document or record must exist, for which role, by when, and what would cause it to fail an audit. It is aimed at compliance and legal teams, engineers, auditors and developers of compliance tools. It is an early release and does not constitute legal advice.
The dataset is structured as a linked graph. Each obligation is a single duty under the Act, written in plain language. It is linked to the article it comes from, the roles it applies to, such as provider or deployer, and the risk categories it covers, such as high-risk or general-purpose AI. Each obligation is also linked to its deadline, the evidence that satisfies it, common audit red flags and mistakes, and primary-source citations.
The current release covers 14 obligations, 17 types of evidence, 11 articles, 2 annexes, 7 deadlines and 6 roles. Deadlines keep a history of changes, such as those introduced by the Digital Omnibus on AI. Each record has a version number and review date. Records still being checked against the legal text are flagged for review. The data is available in JSON and CSV formats, with an interactive audit-readiness checklist, under a CC BY 4.0 licence.
About the tool
You can click on the links to see the associated tools
Tool type(s):
Objective(s):
Impacted stakeholders:
Purpose(s):
Target sector(s):
Country/Territory of origin:
Lifecycle stage(s):
Type of approach:
Maturity:
Usage rights:
Target groups:
Target users:
Stakeholder group:
Validity:
Geographical scope:
People involved:
Required skills:
Technology platforms:
Github stars:
- 1
Use Cases
Would you like to submit a use case for this tool?
If you have used this tool, we would love to know more about your experience.
Add use case
Partnership on AI
