From Sunday, August 2, companies in the European Union will have to let people know when content was created or heavily edited with AI. It’s one of the first big steps in putting the EU’s new AI rules into practice.
The new requirements don’t just apply to one type of content. They cover chatbot responses, AI-made images, audio, and videos, with the idea that people should be able to quickly tell if something came from AI instead of a human.
The disclosure requirement is part of the EU’s broader AI legislation, which is taking effect in stages. The current phase emphasizes transparency by requiring companies to identify content that has been generated or substantially modified by AI. In addition to visible labels for users, the rules also call for technical measures, including watermarks or embedded metadata, to help identify AI-generated material after it is shared.
The new rules arrive as concerns over deepfakes and other AI-generated content continue to grow. As the technology becomes more realistic and widely available, regulators warn that misleading media can spread quickly, making it increasingly difficult for people to tell authentic content from fabricated material.
The new rules mainly target AI content published in professional contexts. If information for the public is created by AI without a human editor reviewing it, it has to be labeled. Personal AI use isn’t included, and the EU has also made exceptions for things like art, satire, fiction, and other creative works.
The regulation requires more than a simple label attached to a post. Companies are expected to use technical measures, including watermarking and digital tagging, that allow AI-generated content to remain identifiable even as it is shared across platforms.
Some of the biggest tech companies are already doing parts of what the EU now requires. TikTok has long asked creators to label AI-generated content and says its tools have tagged more than three billion posts. Meta adds an “AI Info” label to AI-made content on Facebook and Instagram, while Google is working with Nvidia, OpenAI, and Apple on digital tagging standards after signing the EU’s voluntary AI transparency code.
Not everyone thinks the new rules will work perfectly from day one. Google’s Karen Massin says too many labels and disclosures could end up confusing people instead of helping them. If every piece of content carries a warning, users may eventually stop paying attention.
Companies have until December 2 to ensure their existing AI systems comply with the new requirements. Once that deadline passes, regulators are expected to step up enforcement, with companies facing substantial fines for failing to meet the transparency rules.
Maybe you would like other interesting articles?