EU AI Act Self-Assessment

Have you listed all the AI systems and uses in your organisation, including those embedded in third-party tools?

For each use, do you know whether you are a provider or a deployer within the meaning of the AI Act?

Have you checked that none of your uses falls under the prohibited practices (manipulation, social scoring, emotion recognition at work…)?

Have you classified your uses against the AI Act risk levels (high risk, limited risk, minimal)?

For your high-risk uses (HR, credit, education…), have you identified the applicable requirements?

Do you have technical documentation and traceability (logs) for your important AI systems?

Is effective human oversight in place for important decisions assisted by AI?

Do you check the quality and representativeness of the data used by your AI systems?

Are people informed when they interact with an AI (chatbot, assistant)?

Is AI-generated content (text, images, video) labelled as such where required?

Do your terms of use, internal policies or notices describe your AI uses towards customers and employees?

Do you require the necessary information from your AI vendors (documentation, limits, instructions for use)?

Have the teams using or deploying AI been trained on its risks and limits (AI literacy obligation)?

Is it clear who, in the organisation, approves and monitors AI uses?

Do you have a procedure for serious incidents or malfunctions of an AI system?

Do you follow the AI Act implementation timeline to anticipate the deadlines that concern you?