EC Clarifies AI Act Transparency for Deepfakes, Chatbots Before August 2026 Deadline
The European Commission has published draft guidelines detailing the transparency obligations under Article 50 of the AI Act, covering deepfakes, interactive AI, and AI-generated text. These guidelines clarify rules for disclosure, marking of synthetic content, and strict editorial review requirements for AI-generated text related to public interest. Compliance for AI systems and disclosure practices is expected from August 2026, with potential penalties up to EUR 15 million or 3% of worldwide annual turnover for infringements.
Key Takeaways
- Deepfakes must be labeled even without deceptive intent or depicting a real person, applying to any realistic synthetic depiction.
- AI systems require explicit, upfront self-identification at the moment of user interaction, not via fine print or technical terms.
- Editorial review of AI-generated public interest text must be 'genuine and substantive' with clear accountability to qualify for labeling exemptions.
- AI Act obligations apply to interactive AI systems, synthetic content providers, emotion recognition systems, and deployers of deepfakes/public interest AI text.
Why It Matters
The detailed EC guidelines provide critical clarity for streaming platforms and content creators utilizing AI, particularly for generative media and interactive experiences. With the AI Act's transparency obligations set to apply from August 2026, companies must audit existing AI deployments and editorial workflows to ensure compliance. The narrow exceptions and broad definitions signal a strict regulatory environment, necessitating proactive adjustments to avoid significant penalties and potential legal challenges across various content types.
Additional Context
The AI Act, which entered into force on August 1, 2024, will see its transparency rules become fully applicable on August 2, 2026, though specific elements like marking and detection obligations under Article 50(2) might have a revised deadline of December 2, 2026, pending formal adoption of the AI Omnibus proposal (European Commission Digital Strategy, May 2026). Ahead of these deadlines, the European Commission is actively preparing support tools, including a Code of Practice on marking and labeling AI-generated content, which will be a voluntary tool to guide providers and deployers (European Commission Digital Strategy, May 2026). This Code, currently under development by the AI Office with stakeholder input, is expected to finalize by early June 2026 and could offer a streamlined path for demonstrating compliance with Articles 50(2), (4), and (5) of the Act (European Commission Digital Strategy, May 2026). While the guidelines cover Article 50 comprehensively, the Code of Practice will focus on practical technical implementation means for generative AI systems specifically (European Commission Digital Strategy, May 2026). Companies that do not adhere to the Code may face increased evidentiary burdens from market surveillance authorities. The AI Act's risk-based approach highlights transparency as a core risk area, emphasizing informed user decisions when interacting with AI or AI-generated content (European Commission Digital Strategy, May 2026).
Read full article at natlawreview.com
Get this in your inbox → Subscribe
Enjoy our coverage?
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source