Brussels asks platforms to identify content generated by artificial intelligence

Europe 1 with AFP 15:22 pm, June 05, 2023

The European Commission wants online platforms to clearly identify content - images, sounds, texts... - generated by artificial intelligence (AI) to combat the risks of disinformation, announced Monday the vice-president of the European executive, Vera Jourova. The Czech Commissioner for Values and Transparency addressed the request to the signatories of the EU Code of Practice against Disinformation whom she met on Monday in Brussels. This code, non-binding, brings together on a voluntary basis about forty organizations, including major platforms such as Facebook, Google, YouTube, TikTok - while Twitter recently left it.

Obligations for very large platforms from August

"Signatories that integrate generative AI into their services like Bing Chat for Microsoft, Bard for Google should incorporate the necessary safeguards so that these services cannot be used by malicious actors to generate disinformation," Vera Jourova said at a press briefing. "And secondly, signatories who have services that may spread AI-generated disinformation should put in place technology to recognize this content and make it clear to users," Jourova added.

Several fake photographs produced with generative AI have recently been widely shared on social networks, such as that of the pope in a down jacket or the arrest of former US President Donald Trump. Very large platforms with more than 45 million active users in the EU will be subject to obligations under the Digital Services Act (DSA) from 25 August. This regulation provides for "the use of clearly visible marking" to ensure that "an image, audio or video content generated or manipulated (...) is recognizable".

A text that will enter into force "at the earliest end of 2025"

But the Commission wants the label to be introduced "immediately" as part of the code of practice against disinformation. A way to encourage signatory platforms to comply as soon as possible with the DSA, which will apply in a binding manner. The EU is also negotiating specific legislation on AI, which includes transparency obligations for editors such as ChatGPT, but these discussions between member states and MEPs drag on.

If the text were to be adopted before the end of the year, it would enter into force "at the earliest at the end of 2025", according to the European Commissioner for Digital Thierry Breton, who has agreed with the boss of Google, Sundar Pichai, to work by then on rules with voluntary companies.