EU Commission publishes Code of Practice for labeling AI-generated content: Standardized icons and watermarking standards from August 2026
What it really says
The European Commission published the final Code of Practice on marking and labeling AI-generated content on June 10, 2026. The code is a voluntary instrument providing practical measures for providers and deployers of generative AI systems to implement the transparency obligations under Article 50 of the AI Act, which become applicable on August 2, 2026. Key elements include standardized EU icons that creators and publishers can use to label AI-generated content. Technically, the code relies on two mechanisms: digitally signed metadata and imperceptible watermarking, with optional fingerprinting using a registry database. The deepfake definition is deliberately broad: it covers all content that depicts something that could exist in the real world - even without intent to deceive. An exception exists for clearly artistic, creative, satirical, or fictional works, where the disclosure obligation applies in a lighter form. Violations of the underlying legal obligations carry fines of up to 15 million euros or 3 percent of global annual turnover.
Our assessment
For people concerned about deepfakes and AI-generated misinformation, this news is largely reassuring. The EU is creating the first concrete, practically applicable framework for making AI-generated content identifiable. The standardized EU icons are designed to let users see at a glance whether an image, video, or text was created by AI - similar to nutritional labels on food products. The technical requirements for watermarking and metadata go beyond simple labels: they aim to mark AI content in a machine-readable format, enabling platforms and browsers to detect it automatically. However, there is an important caveat: the code of practice itself is voluntary. The underlying obligations from Article 50 of the AI Act are mandatory, but practical implementation depends on whether enough companies sign up and whether enforcement works. The broad deepfake definition - any image that could look realistic - shows that the EU prefers over-labeling to under-labeling. This is positive for consumer protection but could become bureaucratic for creatives and businesses.
Relevance for Germany
The transparency obligations directly affect Germany, as the AI Act applies as EU law and will be enforced from August 2, 2026 - less than two months away. For German companies using generative AI - in marketing, media, or e-commerce - the code provides concrete implementation guidance: they must verify that their AI-generated content is properly labeled and potentially implement technical solutions for watermarking and metadata. For consumers in Germany, the news is particularly relevant because AI-generated content in social media, news portals, and advertising has long been ubiquitous without any standardized labeling. The standardized EU icons could achieve a similar recognition effect as the CE mark on products. The Federal Network Agency, which was just designated as the central AI supervisory authority through the KI-MIG, will also be responsible for monitoring these transparency obligations in Germany.
Fact check
The publication of the final Code of Practice on June 10, 2026 is confirmed by the official European Commission website (Digital Strategy) and the EU Commission's representation in Germany. News agency Agence Europe consistently reports on the publication of the final voluntary Code of Practice. The technical requirements (digitally signed metadata, watermarking) are confirmed by official EU documentation and the IPTC technical analysis. The fine framework of up to 15 million euros or 3 percent of annual turnover corresponds to the AI Act provisions for transparency obligation violations (Article 99 paragraph 4). The effective date of August 2, 2026 for transparency obligations is documented in the official AI Act.
Source
- • https://digital-strategy.ec.europa.eu/en/policies/eu-icons-labelling-ai-generated-content
- • https://digital-strategy.ec.europa.eu/en/faqs/signing-code-practice-transparency-ai-generated-content
- • https://agenceurope.eu/en/bulletin/article/13885/7/european-commission-publishes-final-voluntary-code-of-practice-to-help-label-content-generated-by-ai
- • https://germany.representation.ec.europa.eu/news/kennzeichnung-von-ki-generierten-inhalten-eu-kommission-veroffentlicht-verhaltenskodex-2026-06-10_de