As part of the implementation of the EU’s Artificial Intelligence Act (AI Act), a Code of Practice is being developed to establish guidelines for labeling and disclosing information about content created or modified using AI.
The Code itself is not binding, as noted by AiNews. However, the obligations it outlines are binding. They are governed by Article 50 of the EU AI Act and will apply starting August 2, 2026, regardless of whether a company signs the Commission’s guidelines or not. Signing simply provides companies with a recognized way to confirm their compliance with the requirements.
The new requirements will affect technology developers, online platforms, media companies, and businesses that use generative AI to create digital content.
What changes are European regulations bringing?
The key idea behind the new rules is to make the use of artificial intelligence more transparent for users.
Companies will be required to implement mechanisms to determine whether content was created entirely using AI or significantly altered by such tools.
This primarily concerns:
– images and videos created by generative models;
– synthetic audio;
– text content created by AI;
– deepfakes and other forms of artificially generated content.
European regulators believe that users should be able to understand the origin of digital information, especially in areas where errors or manipulation can have serious consequences—such as politics, the media, finance, and public communications.
The code is intended to help prepare for the AI Act
The Code of Practice on the Labeling of AI Content is part of the broader implementation of the European AI Act.
The AI Act introduces a risk-based approach to technology regulation: different requirements apply depending on the potential impact of the system on citizens’ rights, safety, and public processes.
For companies, this new approach means they must review their internal processes for working with AI.
Businesses will need to consider:
– which artificial intelligence tools are used;
– whether they generate new content or merely assist in processing information;
– whether disclosure of information to users is required;
– How is the origin of digital materials recorded?
The issue of labeling is becoming critical for the market
The rapid spread of generative AI has created a new challenge for the digital economy: it is becoming harder for users to distinguish human-created material from synthetic content.
This applies not only to social media. Similar issues arise in marketing, advertising, journalism, education, financial analysis, and corporate communications.
For companies, transparency regarding AI use may become part of reputation management. Organizations that can explain the origin of their content will gain an additional level of trust from customers and partners.
The changes will have the greatest impact on developers of large language models, image and video generators, as well as digital platforms that distribute user-generated content.
Technology providers will need to implement mechanisms to label AI-generated content or ensure that relevant information about its origin is available.
One possible technical solution is digital content provenance tags, including mechanisms such as watermarks or embedded information about the material’s creation.
However, it remains to be seen how effectively such tools will work given the mass distribution of content and subsequent editing of files.
What This Means for Business and the Media
For companies, the use of AI is gradually shifting from the experimental phase to the realm of regulated activity.
Organizations that use generative AI to prepare advertising materials, analytics, client communications, or internal documents will need to establish their own rules for applying such technologies.
For the media, the issue becomes even more sensitive: the use of AI tools in content preparation requires transparency, quality control, and the preservation of editorial responsibility for published content.
In the long term, the European approach could serve as a benchmark for other markets, as the EU seeks to establish global standards for the regulation of artificial intelligence.
