European Union AI Act mandates machine-readable labeling

The European Union is implementing critical transparency mandates under the EU AI Act. Starting December 2, 2026, providers and operators of generative AI systems, including general-purpose models already on the market, must adhere to machine-readable labeling requirements. This includes disclosing when users are interacting with AI, such as chatbots, and marking AI-generated or modified content with visible icons like ‘AI GENERATED’ or ‘AI MODIFIED’. Deepfakes and AI-generated content regarding sensitive topics like elections, health, or public interest also face explicit labeling obligations.

While certain deadlines have been adjusted, the regulatory burden remains significant. The European Parliament and Council postponed the requirements for standalone high-risk AI systems—covering sectors such as biometric identification, critical infrastructure, and law enforcement—to December 2, 2027. Additionally, AI used as safety components in regulated products, such as medical devices or industrial machinery, will see a deadline extension until August 2, 2028. However, transparency rules under Article 50 are already binding across all 27 EU member states as of August 2, 2026.