The European Commission confirmed on its own site that companies must file to join the Code of Practice on marking AI-generated content by 18:00 CEST on Wednesday 22 July to make the first published signatory list.1 The Code is the voluntary industry standard that sits under the EU AI Act: firms that sign commit to labelling synthetic audio, image and video before Article 50, the Act's binding transparency obligation, applies to new deployments from 2 August.
That cutoff falls on the same day Brussels decides the Paramount-WBD competition review. the Commission finalised the Code on 10 June and confirmed on 7 July that it remains under adequacy assessment. As of 15 July it names no signatories, and no EU broadcaster, streamer or media company has signed.
the Commission tried this same voluntary route with its 2018 Disinformation Code, drew near-zero uptake, and converted it to mandatory standards within roughly two years. The content-marking Code is tracing the same curve, which is why the first broadcaster to sign matters: it sets the disclosure template the rest copy, and dates the moment the industry stopped treating labelling as optional.
Smaller players have already moved. Camb.AI's labelled athletics streams and France Televisions' named-vendor disclosure show compliance is buildable now; the incumbents carrying the most synthetic content are the ones who have signed nothing.
