AI content labelling becomes mandatory across the whole EU starting August 2

Started by Fan, Yesterday at 09:20 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Topic: AI content labelling becomes mandatory across the whole EU starting August 2   Views(Read 56 times)
Active members in this topic:
Fan(1)

Fan

From August 2 a large share of online content made with AI has to carry a label across the EU, this is part of the transparency requirements baked into the AI Act that was adopted two years ago, and it is a much bigger rollout than most of the coverage has made it sound

There are really two separate obligations here, systems that interact with the public using AI like chatbots and virtual avatars must clearly identify themselves as AI, and separately providers have to build in watermarks and other markers so AI generated content can be detected technically even if nobody bothers reading the label

The part that will actually be visible to ordinary people is the deepfake rule, any text, image, video or audio that looks authentic but was generated or manipulated by AI has to be flagged when published as part of someone's professional activity, and even AI written text meant to inform the public on matters of general interest needs a label if there was no real human editorial oversight

Brussels has already published the actual icon templates, simple things like AI, generated by AI, modified by AI, along with a compliance guide, and there are sensible carve outs for artistic, satirical, fictional or clearly creative works, plus a grace period running to December 2 for systems already in operation

Some platforms have basically already done this work voluntarily, TikTok says it has labelled more than 3 billion pieces of content through its creator disclosure tools and detection systems, and Meta already runs an AI Info tag across Instagram and Facebook, so the regulation is partly just formalising practices the biggest platforms already adopted

Not everyone is thrilled though, Google's European affairs lead warned that piling AI labels and legal disclaimers onto everything risks confusing people rather than helping them, and one governance expert compared the whole rollout to the cookie consent banner era, annoying at first, complained about endlessly, and eventually just background noise everyone tunes out

Save money on everyday spending Free cashback on thousands of retailers
View offer