AI watermarking and deepfake labelling will soon be the standard in the EU
As artificial intelligence takes over the internet, the European Union is trying to make sure that we’ll at least be aware it’s happening.
The EU’s landmark risk-based rulebook for tech – the AI Act – puts labelling requirements on AI-generated content. This means companies building AI systems like ChatGPT must ensure that media their tools produce, whether an image or a video, is watermarked as synthetic.
But the law doesn’t only have implications for AI developers. It also says web users should “disclose” if they used an AI system to create a “deepfake” – referring to AI-generated (or enhanced) video, image or audio that could be mistaken for a real person.
Both rules, which kick in later this year, are rather vague. But last week the European Commission released a document that lays out how we might expect them to change the internet in the coming months.
