From this Sunday, AI companies operating in Europe must mark deepfakes and synthetic content as artificial under new obligations in the EU's AI Act — a transparency requirement that civil society groups have demanded for years but experts warn may not be enough to stop the spread of manipulated media. The European Commission developed a voluntary code of practice requiring companies to disclose the artificial origin of synthetic content through machine-readable markings and visible labels, with stricter rules for deepfakes that imitate real people in ways that could deceive viewers.
The rules apply to both companies developing AI technology and those using it professionally, though personal uses are excluded. Major AI companies like OpenAI and Google have broadly backed the transparency requirements by signing the Commission's code of practice, while warning that detection and marking technologies remain a moving target.
What Counts as a Deepfake
Brussels defined deepfakes as realistic AI-generated or AI-manipulated images, audio or videos that imitate a real person's appearance, voice or actions in a way that could lead people to believe they said or did something that never happened. Not all synthetic content is deceptive, the Commission said. Content created for clearly artistic, creative, satirical or fictional purposes generally falls outside the disclosure requirements.
The distinction matters. A 2023 AI-generated image of Pope Francis wearing a fashionable white puffer jacket went viral as satire. But 2024 AI-generated explicit images targeting Taylor Swift spread online without her consent, and a 2022 fake video showed Ukrainian President Volodymyr Zelenskyy telling soldiers to surrender shortly after Russia launched its full-scale invasion of Ukraine. The latter examples represent the kind of manipulated media the EU is targeting — content designed to deceive, harm or undermine democratic discourse.
The Detection Problem
The code of practice also addresses detection mechanisms and requires collaboration among AI companies, social media platforms, civil society organisations and fact-checkers. But there's currently no single industry-wide solution. Companies are experimenting with watermarking, metadata and content provenance systems that don't always work together or talk to each other.
Experts warned that watermarks can be removed, altered or lost when content is edited, compressed or shared across platforms. Researchers argued that no single technology will be enough. The European Commission's technical study concluded that a combination of solutions is more powerful than any single measure and that a best practice is to implement multiple layers of marking and detection for AI-generated content.
Implementation Challenges
The EU adopted sweeping rules to regulate artificial intelligence two years ago and created a dedicated AI Office to oversee implementation. The transparency requirements now taking effect represent one pillar of that broader regulatory framework — but enforcement will depend on coordination across borders, platforms and sectors. Civil society groups have long called for stronger protections against manipulated media, particularly as synthetic content becomes easier to produce and harder to distinguish from reality.
The rules won't stop all deepfakes. But they do establish a legal baseline: companies that fail to mark AI-generated content as artificial will face regulatory consequences. Whether that's enough to protect democratic discourse, individual dignity and public trust in media remains an open question.
Why This Matters:
Deepfakes aren't just a technical problem — they're a threat to democratic accountability, individual safety and the shared reality that makes public debate possible. The fake Zelenskyy video showed how synthetic media can be weaponised in wartime. The Taylor Swift images showed how women in particular are targeted with non-consensual explicit content. And the Pope Francis jacket showed how easily satire can be mistaken for reality when AI-generated images circulate without labels. The EU's transparency rules are a necessary first step, but experts warn that marking content as artificial won't be enough if the marks can be stripped away or ignored. The real test will be whether platforms, AI companies, civil society and regulators can coordinate detection and enforcement at scale — and whether Europe's regulatory model can set a global standard before the technology outpaces the law.