TIKTOK is finally letting users cut down on AI content with a brand new slider.
The social media giant also revealed that it’s now adding an “invisible” red flag for AI content that regular users can’t see – allowing TikTok to track dodgy fake clips even if they’ve been re-edited elsewhere.
AI apps can now churn out convincing videos quickly and cheaply.
And much of that is making its way onto social media apps, including TikTok.
Now TikTok has confirmed that it is adding new ways to identify AI, and giving you the power to shrink its presence in your For You feed.
A new slider will soon appear under TikTok’s Manage Topics settings, which is already an existing feature.
Read more on social media
FACE VALUE
Facebook brings BACK popular money-making feature two years after it was closed
PAY APP!
Facebook & Instagram to charge £4 a month for no ads – but trick makes it cheaper
Normally this page lets you adjust how often you see content across common categories like Dance, Sports, and Food & Drinks.
But there’s a new AIGC (“AI Generated Content”) slider that lets you “dial things down”.
“At TikTok, we believe AI can transform how people share their creativity, discover new passions, and stay safe on our platform, when used transparently and responsibly,” TikTok said at a special event around user safety at its Dublin offices.
“That’s why we invest in AI-powered experiences that create value for our community, while building transparency tools to spot, shape, and understand AI-generated content.”
There’s no way to turn off AI content on TikTok completely, The Sun understands.
But TikTok told The Sun that it will give your For You Page algorithm a “nudge” to limit AI content if you drag the new slider to the left.
And you can still use the “Not Interested” button when you see an AI video to help the TikTok algorithm learn that you don’t want to see that kind of content.
Dragging the slider should be your main port of call for reducing AI content in your feed, however.
“This type of content typically includes posts created or edited with AI,” TikTok’s Manage Topics pop-up explains.
“Adjusting this may apply to all topics.”
To spot and flag AI content, TikTok is using two main systems.
The first is a “cross-industry technology” called C2PA Content Credentials.
These are like flags that get embedded into a video’s metadata – the pack of info that comes along with a photo or footage on the internet.
TikTok says that using this system has helped to label more than 1.3 billion videos so far.
But TikTok warned that it’s not a foolproof system.
“A common industry challenge is that these kinds of labels may be removed when content is reuploaded or edited elsewhere,” TikTok warned.
That’s where the second trick comes in: “Invisible watermarks”.
Traditionally, a watermark is a flag that visibly appears on an image to show its creator.
TIKTOK – THE LATEST NUMBERS
Here are some fresh TikTok stats straight from the social media giant…
Usage:
- Over a billion TikTok users globally
- More than 200 million TikTok users in Europe
- Over 100 million pieces of content uploaded to TikTok daily
Moderation:
- 96% of offending content removed proactively before going live
- 86% of offending content now removed through automation
- More than 400 million rule-breaking videos removed in first half of this year (less than 1% of overall videos posted)
- Over a billion rule-breaking comments removed in first quarter of 2025
OpenAI has been using visible watermarks on videos generated by its Sora AI, which you’ll see floating around on top of the clip.
But savvy editors can remove these in a bid to conceal the clip’s AI origins.
TikTok says it’ll be rolling out “invisible watermarks”, initially for content created or edited with the company’s own AI tools in the app.
“The common industry challenge is that with these methods, the labels might get removed if you download the content, re-upload it, or re-edit it elsewhere,” said Jade Nester, TikTok’s Director for Data Public Policy in Europe..
“So these invisible watermarks help us address this by adding a robust technological watermark that only we can read, making it harder for other people to remove that.
“And we can use this watermark to label content more reliably.
SEE CLEARLY
Mechanic reveals ‘secret’ button that will defrost your windows even quicker
waltzing off
La Voix’s statement in full as injured star is forced to pull out of Strictly
“We’ll also getting more context around any changes that the content went through.
“Over the coming weeks, we’ll begin testing the invisible watermarks with AI generated content made with certain AI TikTok tools.”
