On 2 December 2026, two specific AI Act changes are scheduled to apply under the current amended timeline. New prohibitions in Article 5(1)(ba) and Article 5(1)(bb), added by Regulation (EU) 2026/1744, concerning AI systems used to generate or manipulate non-consensual intimate material and child sexual abuse material take effect, and the transition period for certain providers of systems already placed on the market before 2 August 2026 ends for the Article 50(2) marking and detection obligation. This is a focused milestone, not a second general application date for the whole Act.
Why 2 December 2026 Matters
The December 2026 milestone is narrower than the broad 2 August 2026 application date. Most organisations will already have been working with applicable rules on AI literacy, prohibited practices, general-purpose AI models and transparency for several months. The December date adds two specific changes identified in the official implementation timeline: additional prohibitions involving certain forms of harmful synthetic content, added by Regulation (EU) 2026/1744, and the end of a transition for the Article 50(2) provider obligation for some systems that were on the market before 2 August 2026. Teams should therefore treat this as a targeted update to an existing compliance programme rather than a fresh start.
New Prohibited AI Practices Begin Applying
The new prohibitions are Article 5(1)(ba) and Article 5(1)(bb), and they concern AI systems used to generate or manipulate specified non-consensual intimate material and child sexual abuse material, described here in general terms only. Because prohibited-practice analysis is use-case specific, organisations should examine what the system is intended to do, how it is actually used, who controls the generation or manipulation process and what safeguards or restrictions exist. A broad label such as 'image generator' or 'content tool' is not enough to determine whether the prohibition is engaged. Providers and deployers should update prohibited-practice screening so the December 2026 additions are considered alongside the prohibitions that have applied since February 2025.
For the broader regulatory framework, see EU AI Act timeline.
The Article 50(2) Transition Ends
Article 50(2) concerns providers of AI systems that generate synthetic audio, image, video or text content and requires outputs to be marked in a machine-readable format and detectable as artificially generated or manipulated, subject to the provision's conditions and exceptions. The official timeline provides a transition until 2 December 2026 for certain systems, including some GPAI-based systems, that were already placed on the market before 2 August 2026. Providers relying on that transition should be able to identify the relevant system version, market date and technical implementation work rather than assuming all existing products receive the same extension.
Provider Marking and Deployer Disclosure Are Different
The provider marking obligation in Article 50(2) should be kept separate from disclosure duties that can fall on deployers under other parts of Article 50. A provider may need to design technical marking or detection into generated content, while a deployer may need to disclose certain uses such as deepfakes or specified public-interest text. The responsible actor, timing and form of the duty therefore differ. A compliance record should identify which paragraph was assessed, which operator carries the duty and how the requirement is implemented in the real product or publishing workflow. One generic 'AI-generated' label does not answer every Article 50 question.
What Evidence Should Affected Providers Keep?
When the transition ends, affected providers should be prepared to demonstrate how their marking and detection approach works in the relevant system. Useful evidence can include the product version, output types covered, technical method, testing results, known limitations, exception analysis and change-control records. Organisations using third-party models or generation services should also document what marking information is supplied by the upstream provider and what survives downstream processing. The objective is not to create paperwork for its own sake, but to preserve a traceable explanation of how the applicable requirement was interpreted and implemented for the actual system.
Continue with what changes on 2 December 2027.
Which Deadlines Still Come Later?
The December 2026 milestone does not bring forward the main high-risk application dates. Annex III high-risk system rules remain scheduled for 2 December 2027, and high-risk rules for systems embedded in regulated products under Annex I remain scheduled for 2 August 2028, under the timeline as amended by Regulation (EU) 2026/1744. Member States should have at least one AI regulatory sandbox operational by 2 August 2027, in the Service Desk wording, which is a separate milestone between the December dates. Nor does the December transition postpone transparency duties generally: Article 50 started applying on 2 August 2026. Organisations should therefore maintain separate date fields for each obligation. A single 'AI Act deadline' column can hide the fact that one system may already face transparency requirements while its high-risk obligations are still in preparation.
How to Prepare Before 2 December 2026
Before 2 December 2026, providers should confirm whether any system was placed on the market before 2 August 2026 and whether the Article 50(2) transition is actually relevant. Product and engineering teams should verify the marking and detection implementation, while legal and governance teams update prohibited-practice screening for the new synthetic-content prohibitions in Article 5(1)(ba) and Article 5(1)(bb). Deployers should separately review their own disclosure duties. Record owners, test results, exceptions and unresolved questions, then schedule a post-deadline review so later changes to models, output pipelines or intended uses do not silently invalidate the assessment.
Frequently asked questions
Does every AI system face a new requirement on 2 December 2026?
No. This is a targeted milestone concerning new prohibited practices in Article 5(1)(ba) and Article 5(1)(bb) and the end of a specific Article 50(2) transition for certain systems already on the market.
Does the Article 50(2) transition postpone all transparency duties until December 2026?
No. Article 50 generally began applying on 2 August 2026. The December date concerns a limited transition for the marking and detection obligation for certain pre-existing systems.
Do Annex III high-risk rules start on 2 December 2026?
No. Under the current amended timeline, Annex III high-risk rules apply from 2 December 2027, as amended by Regulation (EU) 2026/1744.
What does this article not decide?
This article does not decide whether any particular system engages the new prohibitions or benefits from the Article 50(2) transition. Those questions turn on the system's function, market date and the exact legal wording, and should be confirmed with a competent authority or legal adviser where they are contested.
Continue through this AI Act cluster
This article is an informational overview, not a legal determination. Check the current legal text and official implementation guidance for your system and role.
Consolidated Regulation (EU) 2024/1689, as amended ↗AI Act Service Desk implementation timeline ↗AI Act Service Desk: Article 50 ↗