What Exactly Is Changing?
Starting August 2, 2026, a key part of the EU's comprehensive AI Act came into force, specifically targeting transparency. Under these new rules, any content that is artificially generated or manipulated to look authentic must be clearly labelled. This
applies to a wide range of synthetic media, including AI-generated images, video, audio, and even text on matters of public interest that lacks human editorial oversight. The goal is to ensure people know when they are looking at a deepfake, interacting with a chatbot, or reading an article written by a machine. The rules require both a visible mark on the content itself and a machine-readable digital watermark to prove its artificial origins.
Why Was This Rule Introduced?
The rapid rise of sophisticated AI tools has made it easy to create convincing but fake content, from false audio of politicians to realistic images of events that never happened. Regulators and lawmakers are concerned that this flood of synthetic media could be used to spread disinformation, interfere with elections, and erode public trust. A European Parliament member involved in the negotiations described the transparency measure as crucial for protecting democracy and the authenticity of facts online. By making it clear what is real and what is artificial, the EU hopes to empower citizens and reduce the potential for manipulation and deception.
Who Must Comply?
The obligations apply broadly, affecting both the 'providers' who create AI systems and the 'deployers' who use them. This means companies building generative AI tools (like Google or OpenAI) must design them to embed machine-readable watermarks. Meanwhile, businesses, agencies, and even individuals who use these tools to publish content commercially in the EU are responsible for ensuring it is properly labelled. The rules apply to any content consumed within the EU, regardless of where the creator is based. However, there are exceptions; the labelling mandate does not apply to personal, non-commercial content or to works that are “evidently artistic, creative, satirical or fictional”.
How Will The Labels Work?
Companies and creators have some flexibility. The European Commission has provided a set of freely available black-and-white icons that can be used to label AI content, but companies are also permitted to design their own. The key requirement is that the label must be clearly visible to a user upon their first encounter with the content. Beyond the visible label, providers of AI systems must also embed machine-readable signals, like digital watermarks, that allow platforms to automatically detect synthetic content. Major platforms like YouTube, Instagram, and TikTok have already started implementing their own labelling systems in anticipation of these rules.
What Are the Penalties for Non-Compliance?
The EU is backing these transparency rules with significant financial penalties. Companies that fail to comply with the labelling obligations can face fines of up to €15 million or 3% of their total worldwide annual turnover, whichever is higher. For the most serious violations of the overall AI Act, such as using prohibited AI practices like social scoring, fines can reach as high as €35 million or 7% of global turnover. The enforcement will be carried out by the European Commission's AI Office in conjunction with national authorities in member states.
What This Means for India and the World
While these are EU regulations, their impact will be felt globally. Because the rules apply to any content viewed by users in the EU, creators and companies in India and elsewhere will need to adapt if they serve a European audience. This phenomenon, known as the 'Brussels effect,' often sees EU regulations become de facto international standards as global companies find it easier to apply one set of rules across all their operations. With major tech platforms already rolling out labelling features worldwide, users in India will likely see more AI disclosures on their feeds. The EU's move also sets a major precedent that could influence future AI legislation in other countries, including India, as governments worldwide grapple with how to govern this powerful new technology.
