{"page":"/blog/eu-ai-act-advertising","modified":"2026-11-10","license":"https://creativecommons.org/licenses/by/4.0/","columns":["Provision","What it requires","Applies from","Relevance to ad tech"],"rows":[["Article 2(1)(a) to (c), scope","Covers providers placing AI on the EU market wherever established, deployers in the EU, and non-EU providers and deployers \"where the output produced by the AI system is used in the Union\"","In force 1 August 2024","Non-EU DSPs, creative tools and agents whose output reaches EU users"],["Article 4, AI literacy","Providers and deployers \"take measures to support the development of AI literacy\" of their staff; wording replaced by 2026/1744","2 February 2025; new wording from 27 July 2026","All four uses: bidding models, creative tools, agents, measurement"],["Article 5(1)(a), manipulation","Bans subliminal, purposefully manipulative or deceptive techniques that materially distort behaviour and cause, or are reasonably likely to cause, significant harm","2 February 2025","Targeting and creative, only where every condition is met"],["Article 5(1)(b), vulnerabilities","Bans exploiting vulnerabilities due to age, disability or a specific social or economic situation, with the same harm test","2 February 2025","Targeting of children, people in debt or other vulnerable groups"],["Article 5(1)(c), social scoring","Bans scoring people on social behaviour or traits that leads to unjustified or unrelated detrimental treatment","2 February 2025","Audience scores reused to deny offers in unrelated contexts"],["Article 5(1)(f), emotion recognition","Bans inferring emotions in workplaces and education, except for medical or safety reasons","2 February 2025","Measurement tools that read emotions in those settings"],["Article 5(1)(g), biometric categorisation","Bans categorising people from biometric data to infer race, political opinions, religion, sex life or orientation, among others","2 February 2025","Targeting built on face or voice analysis"],["Article 5(1)(ba) and (bb), added by 2026/1744","Bans AI that generates non-consensual intimate images of identifiable people or child sexual abuse material","2 December 2026","Creative generation tools and their safeguards"],["Article 6(2) and Annex III","High-risk requirements in Chapter III, Sections 1 to 3","2 December 2027","Job ads in recruitment, credit scoring, life and health insurance pricing, election influence"],["Article 6(3)","An Annex III system is not high-risk if it poses no significant risk, but \"shall always be considered to be high-risk where the AI system performs profiling\"","With Annex III, 2 December 2027","Profiling models used inside an Annex III area"],["Article 50(1)","Providers design systems that talk to people so users are told they are interacting with AI, unless obvious","2 August 2026","Agents, chatbots and conversational ad units"],["Article 50(2)","Providers of generative systems mark outputs in a machine-readable format","2 August 2026; 2 December 2026 for systems on the market before 2 August 2026","Creative generation vendors and in-house tools"],["Article 50(3)","Deployers of emotion recognition or biometric categorisation inform the people exposed","2 August 2026","Measurement that reads faces or emotions, where not banned"],["Article 50(4)","Deployers disclose deep fakes, and AI text published to inform the public on matters of public interest","2 August 2026","AI ads showing realistic people, products or places"],["Article 53","General-purpose model providers keep documentation, inform downstream providers, adopt a copyright policy and publish a training summary","2 August 2025; 2 August 2027 for models on the market before 2 August 2025","Ad tech firms building on large models receive Article 53(1)(b) documentation"],["Article 99","Fines up to EUR 35 million or 7% for Article 5; up to EUR 15 million or 3% for Article 50 and deployer duties; up to EUR 7.5 million or 1% for misleading information","Penalty chapter from 2 August 2025","All uses; SMEs and small mid-caps pay the lower of the two amounts"]],"sourceUrls":[[],[],[],[],[],[],[],[],[],[],[],[],[],[],[],[]]}