The European Union’s Artificial Intelligence Act (AI Act) introduced its first set of obligations on February 2, 2025. This landmark legislation establishes a phased framework for governing AI technologies across the EU.
- Ban on certain AI practices: The AI Act prohibits AI systems that exploit vulnerable groups or conduct manipulative social scoring and biometric identification in public spaces.
- Establishment of governing bodies: It creates the AI Office and AI Board to oversee implementation and enforcement, ensuring consistent application across member states.
- Member State responsibilities: EU countries must designate surveillance and notifying authorities to monitor compliance and report infractions.
- General-purpose AI model obligations: Providers of general-purpose AI (GPAI) models must prepare detailed technical documentation, provide summaries of training data sources and ensure that training respects copyright laws.
These initial obligations mark the beginning of the EU’s phased regulatory approach. Further requirements, including obligations for models with systemic risks and additional governance measures, will come into effect in August 2026.
