The EU AI Act: what developers and adopters should know

Different rules for different risks

The EU AI Act does not treat every AI system alike. The regulation distinguishes prohibited practices, transparency duties, rules for general-purpose AI models, and more extensive requirements for certain high-risk systems.

A company should first ask not merely “do we use AI?” but which system it uses, for what purpose, and in which legal role: provider, distributor, or deployer. Classification affects the applicable duties.

Dates matter

The regulation entered into force on 1 August 2024, with obligations applying in stages. General provisions and some prohibitions applied from February 2025; rules for general-purpose AI models from August 2025. A significant set of provisions, including transparency obligations, began applying on 2 August 2026. Following changes to the timetable, some high-risk obligations apply from December 2027 or August 2028, depending on the system category.

A practical starting point

Organizations should inventory AI systems, document use cases, determine classifications, identify their role, and keep proportionate records. Staff literacy, human oversight, and incident procedures are also relevant.

Requirements differ by sector and product. Compliance decisions should rely on the current legal text and qualified advice where appropriate. This article is general information, not legal advice.