- From August 2, AI firms must label deepfakes and AI-generated content as artificial under EU AI Act.
- People should know AI origin; realistic deepfakes risk deception, harm and political manipulation online.
- Experts warn that technical gaps and a lack of common standards could undermine enforcement of the rules.
The European Union (EU) is forcing tech firms and artificial intelligence (AI) professionals to label AI-generated content from Sunday, August 2, 2026, under the EU’s AI Act. The new transparency rules require clear watermarks and disclosures on synthetic content to help users identify what is real and what is AI-generated.
EU AI Act Mandates Labels for AI-Generated Content From August 2
From 2 August 2026, key transparency obligations under Article 50 of the EU AI Act (Regulation (EU) 2024/1689) will take effect. Key requirements are synthetic content labeling, machine-readable marking, chatbot disclosure and public interest text labeling.The obligations cover a wide range of systems on the EU market. This does not apply to personal, non-professional use.
To facilitate compliance, the European Commission released final guidelines on July 20, 2026, and a voluntary Code of Practice on Transparency of AI-Generated Content. These include optional standardized icons for consistent labelling, such as “AI GENERATED” or “AI MODIFIED”.
Why the EU Mandates AI Labeling to Combat Deepfakes
Deepfakes are hyper-realistic AI-generated or AI-manipulated photos, audio or video of a person that makes them look like they are saying or doing something they never actually did. They present clear risks of political manipulation, fraud, reputational damage and spreading disinformation.
For example, the risks have been highlighted by high-profile cases, such as a fake video of Ukrainian President Volodymyr Zelenskyy encouraging troops to surrender in 2022, and the non-consensual use of AI to generate explicit photos of Taylor Swift in 2024.
EU lawmakers concluded that increased transparency is the most proportional reaction to the ever increasing risk of synthetic content. The ultimate goal is not to ban AI-generated content, but to make sure that Europeans can continue to trust the information environment that underpins their democracy and social cohesion.
What’s Next as Experts Warn of AI Enforcement Gaps
As the EU’s labelling rules take effect this Sunday, experts and industry players are already flagging key enforcement gaps that could limit their real-world impact. Geographic reach is the biggest hurdle, since deepfakes developed elsewhere can easily and rapidly reach millions of users in the EU, yet only firms active in or targeting the EU market are covered by the AI Act.
Another significant challenge is still technical limitations. While companies like OpenAI and Google have implemented steps towards transparency, experts note that detection and marking technologies are still developing. There are no universal standards as tech companies are using different approaches, including digital watermarking, metadata, and content provenance systems
In addition, experts caution, digital markers can be removed, altered or lost during editing or sharing of content. The EU is therefore aiming for a multi-layered approach with several detection techniques. Authorities will be monitoring compliance and the fines for breaches could reach up to €15 million or 3% of global turnover.
Related: Europe’s ESMA Launches CSA to Increase Crypto Compliance
Disclaimer: The information presented in this article is for informational and educational purposes only. The article does not constitute financial advice or advice of any kind. Coin Edition is not responsible for any losses incurred as a result of the utilization of content, products, or services mentioned. Readers are advised to exercise caution before taking any action related to the company.
