EU AI Act: Deepfake Labeling Becomes Mandatory August 2
What it really says
On August 2, 2026, just 13 days away, the transparency obligations under Article 50 of the EU AI Act take effect. The four core obligations: First, AI systems that interact with humans (chatbots, virtual assistants) must clearly identify themselves as AI unless this is already obvious. Second, providers of generative AI must ensure that synthetic content (audio, image, video, text) is both machine-readably marked and detectable as AI-generated. Third, deepfakes, meaning AI-generated or manipulated image, audio, or video content, must be labeled as artificially produced. Fourth, companies must disclose when they use AI for emotion recognition or biometric categorization. The transparency obligations apply not just to high-risk systems but to all AI systems in these four use cases. Non-compliance can result in fines of up to 15 million euros or 3 percent of global annual revenue. Generative AI systems already on the market before August 2 have until December 2, 2026 to meet the technical marking requirement under Article 50(2), per the Digital Omnibus. The stricter rules for high-risk AI systems (such as in HR or healthcare) have been deferred to December 2027 and August 2028 respectively.
Our assessment
Starting August 2, chatbots in the EU must disclose that they are not human, and deepfakes must be labeled as AI-generated. For consumers, this is good news: the transparency obligations create a basis for distinguishing AI-generated content from human content. The deepfake labeling requirement in particular addresses a growing problem, as synthetic media are increasingly used for disinformation and fraud. Enforcement will be a challenge, however: those who intentionally use deepfakes for deception are unlikely to voluntarily comply with labeling requirements. The real impact will be on legitimate providers and platforms that must now implement standards. The deferral of high-risk obligations to 2027/2028 is worth scrutinizing, as AI in HR and healthcare carries the greatest risks. The transparency rules are a necessary first step but not sufficient on their own.
Relevance for Germany
This deadline directly affects German companies. According to surveys, 56 percent of German companies see significant need to catch up on implementation, and 74 percent of firms with more than 250 employees have not yet completed preparations. This means every company operating a chatbot on its website must ensure within 13 days that it identifies itself as AI. Marketing departments using AI-generated images or videos must label them accordingly. The Munich Chamber of Commerce (IHK) and TUV have already published guidelines, and anwalt.de offers recommendations for SMEs. For German consumers, the regulation means greater everyday transparency: when a customer service chat, product description, or social media content originates from AI, this must now be disclosed. Enforcement responsibility in Germany lies with the Federal Network Agency (Bundesnetzagentur) as the national AI supervisory authority.
Fact check
The effective date of transparency obligations under Article 50 on August 2, 2026 is established in the EU AI Act itself and confirmed by numerous legal analyses (Sidley Austin, Pearl Cohen, William Fry, Lexology). The four core obligations (chatbot disclosure, content marking, deepfake labeling, emotion recognition) are taken directly from the Article 50 legal text. The fine ceiling of up to 15 million euros or 3 percent of revenue is set in Article 99 of the EU AI Act. The transitional period for existing generative AI systems until December 2026 comes from the Digital Omnibus signed on July 8, 2026. Survey data on German companies' lack of preparation comes from reporting by Sage and TUV Consulting. The Bundesnetzagentur's role as national AI supervisory authority is established by the AI Supervision Act.
Source
- • https://artificialintelligenceact.eu/article/50/
- • https://artificialintelligenceact.eu/transparency-rules-article-50/
- • https://datamatters.sidley.com/2026/06/24/eu-ai-act-transparency-obligations-preparing-for-compliance-by-2-august-2026/
- • https://www.anwalt.de/rechtstipps/eu-ki-verordnung-2026-was-unternehmen-bis-2-august-umsetzen-muessen-271804.html