TikTok, an app that was as soon as only a place for user-generated content material, is launching a brand new setting that lets customers select how a lot AI-generated content material they need to see of their “For You” feed. The corporate can be introducing extra superior labeling applied sciences for AI-generated content material.
The brand new AI-generated content material (AIGC) management is rolling out throughout the app’s “Handle Subjects” instrument, which lets customers select what they see on TikTok.
“Handle Subjects already allows individuals to regulate how typically they see content material associated to over 10 classes like Dance, Sports activities, and Meals & Drinks,” TikTok defined in a blog post. “Like these controls, the AIGC setting is meant to assist individuals tailor the various vary of content material of their feed, reasonably than eradicating or changing content material in feeds totally.”
The transfer comes as firms like OpenAI and Meta are embracing AI-only feeds. In September, Meta launched Vibes, a brand new feed for sharing and creating brief AI-generated movies. A couple of days after Meta’s launch, OpenAI launched Sora, a social media platform for creating and sharing AI-generated movies.
Since Sora’s launch, real looking AI-generated movies have been posted to TikTok. Moreover, many TikTok customers are leveraging AI to create visuals for posts about different subjects, like history or celebrities.
TikTok says that with the brand new AI-generated content material management, customers who need to see much less of this kind of content material can now dial issues down, whereas those that get pleasure from it might probably select to see extra of it.
You may entry the brand new functionality by going into your Settings, deciding on “Content material Preferences” after which clicking the “Handle Subjects” possibility. Then, you may transfer the slider for various subjects, together with AI-generated content material, to regulate how a lot you do or don’t need to see that kind of content material in your For You feed.
The change is rolling out within the coming weeks, TikTok says.
To enhance its means to label AI-generated content material, TikTok is now testing a expertise known as “invisible watermarking.”
TikTok already requires individuals to label real looking AI-generated content material and makes use of a cross-industry expertise known as Content Credentials from C2PA, which embeds metadata into content material that lets it and different platforms know when one thing is AI-generated. Nevertheless, TikTok notes that these labels may be eliminated when content material is reuploaded or edited on different platforms.
With the brand new “invisible watermarks,” TikTok will add one other layer of safeguards by utilizing a watermark that solely it might probably learn. Meaning it’ll be more durable for others to take away it.
TikTok will begin including invisible watermarks to AI-generated content material made with TikTok instruments like AI Editor Professional. It’s additionally including them to content material uploaded with C2PA’s Content material Credentials. The corporate says these watermarks will assist it label content material extra reliably. TikTok notes that it’ll proceed studying C2PA’s Content material Credentials and add them to AI-generated content material made on its platform.
Associated to those efforts, TikTok additionally introduced that it’s launching a $2 million AI literacy fund geared toward consultants, just like the nonprofit Girls Who Code, to create content material that teaches individuals about AI literacy and security.
