TikTok will provide AI-generated external content with a warning label – IT Pro – News

TikTok will provide AI-generated external content with a warning label – IT Pro – News
TikTok will provide AI-generated external content with a warning label – IT Pro – News
--

TikTok will start flagging videos and images created by artificial intelligence. The social network adds a watermark. This already applied to AI content created in TikTok, but will now also apply to content from outside the app.

TikTok says it initially plans to add watermarks to images and videos uploaded to TikTok. In the future, this watermark will also be available for audio content. That watermark also remains in images and videos when users download them from the app, so that new users on other social networks can also see that it is artificially generated content when a user shares it.

That process is done via Content Credentials, a standard set up by the Coalition for Content Provenance and Authenticity. This C2PA is a partnership of dozens of tech companies, including Adobe, Google, Microsoft and OpenAI. The coalition maintains the Content Credentials standard. Companies offering generative AI services, such as Adobe in Photoshop or OpenAI in ChatGPT, have implemented this as standard. Then, when a user generates an image or video, metadata is automatically added to it, in addition to a watermark.

When TikTok then sees content with such metadata added, it will show that label. Users can then click on that to see the metadata and learn more about where the image came from.

TikTok is the first major social network to implement the standard in this way. The company therefore warns that initially not much content will appear showing the watermark, because many generative AI services have not yet fully implemented the standard.

TikTok does not give an explicit reason why it is now starting with such content. It is clear that social media are under pressure from policymakers to take tougher action against AI content on their platforms. Many policymakers and experts fear that fake news and disinformation are being spread in the run-up to the European elections next month and the US presidential elections in November. This can be enhanced by AI.


The article is in Dutch

Tags: TikTok provide AIgenerated external content warning label Pro News

-

PREV Crypto Analyst: Altcoins Hit Bottom in June, Then Start Bull-Run
NEXT Investor behind owner of Babboe bicycles must write off tens of millions | RTL News