With the amount of AI-generated content steadily rising, social networks are trying to get ahead of the wave, while also trying to ride it with their own generative AI features.
TikTok is the latest app to launch new rules around AI disclosure, with the platform now notifying users that they’ll need to tag AI-generated content in the app, or risk having it removed by moderators.
As you can see in this screenshot, shared by social media expert Matt Navarra, TikTok is now directly prompting users to tag AI-generated content that depicts “realistic scenes”.
Which is a somewhat vague descriptor, but essentially, TikTok’s trying to get ahead of the expected wave of AI deepfakes, by enforcing its rules around related disclosure.
TikTok actually updated its rules on this element back in March, but it’s now stepping up its enforcement, with a dedicated tag for AI-generated content that creators can activate within the upload flow.
TikTok’s the latest platform to add a specific AI tag, with Instagram also developing its own AI content labels, which it too will likely make a posting requirement. YouTube’s also developing new tools to deal with the expected “AI tsunami”, while X, thus far, has put its reliance on Community Notes to help keep users informed of artificial content.
TikTok is actually a step ahead in this respect, because parent company ByteDance already added AI labels to the Chinese version of the platform earlier this year. So TikTok can lean into that experience, and ensure that it’s meeting the full requirements of such at launch.
Which is a key benefit that TikTok has in general. A lot of new features in the app look polished and clean at launch, because most of them have actually been available to Chinese users for months before they make the transition across. Which is why TikTok’s best placed to stay a step ahead. So if you want to know what’s coming to the app, take a look at the latest Douyin updates.
And this will be a key element, for all social apps. With tools like DALL-E and Midjourney becoming more and more commonly used, it’s inevitable that we’re going to be confronted with more AI-generated content, and not all of it will be as easy to spott as Harry Potter characters wearing Balenciaga.
During major conflicts, election cycles, natural disasters, social media platforms play a key role in keeping people informed, and there’ll undoubtedly be more instances where people are confused about what their eyes are seeing.
If only we could add the same labels to ChatGPT-originated text as well, so we could catch out the engagement farming junk as it’s posted.