Entertainment app TikTok is setting new standards in digital transparency by introducing a rating of AI-generated content on its platform. This initiative aims to increase authenticity and provide users with clarity about the origin of content.
It is becoming increasingly difficult to distinguish between artificially generated content and real content, and many social platforms see it as their duty to ensure transparency and authenticity on their own platforms. The Meta Group, which owns Facebook, Instagram, and Threads, recently announced that it will begin labeling AI-generated images with its own watermarks in the coming months. YouTube also recently announced a new transparency tool that allows creators to label their AI-generated videos themselves.
Back in September 2023, TikTok introduced new meta labels as part of a policy change for AI content. Creators are encouraged to tag AI-generated content using this tool. Meanwhile, TikTok has begun tests for automatic AI content classification.
It appears that the tests were successful because The entertainment app was recently announced on its own blogIt will automatically start classifying AI-generated content such as photos, videos, and audio files posted by users on the platform.
Transparency through content credentials
In cooperation with Alliance for Content Source and Authenticity (C2PA), co-founded by Adobe, provides TikTok content credentials. This technical solution works similarly to a “feed card” for digital content by associating metadata with files that contain details about where and how the content was created or edited. This allows users to verify the origins and editing history of content, thus promoting greater credibility.
We are very pleased that TikTok, with its huge community of creators and users, has decided to work with C2PA and CAI to ensure more transparency and authenticity on the platform. At a time when digital content can change rapidly, it is important to provide ways to differentiate between AI-generated content and raw content. Today’s TikTok announcement is an important step in our journey to achieve this goal.
He says Dana Raogeneral counsel and chief trust officer at Adobe, and co-founder of C2PA.
TikTok takes AI classification one step further by implementing a system that uses cryptoasset hashing to ensure content is tamper-proof. This means that any change in content or attached data can be detected, representing an important step forward in security and trust.
Content credentials remain visible after downloading
Content credentials for AI-generated TikTok content will remain viewable even after the content is downloaded. Users can then use C2PA's verification tool to identify AI-generated content as such and even see when, where and how content on TikTok was created and if it has been edited. Other platforms that support content credentials can automatically use them for naming.
However, as technology advances and provides more and more transparency to users, questions remain, especially regarding the detailed description of changes made to the image. Because the AI filter to brighten the image using AI classification is equivalent to the image that was completely generated using AI.
Sell or ban?
TikTok files a lawsuit against US law
“Certified tv guru. Reader. Professional writer. Avid introvert. Extreme pop culture buff.”
More Stories
H145 with five-blade rotor put into service / Christoph 100 to start with new technology in future
5G technology only from 2026 – Allgemeine Zeitung
Climate Policy: Technology Alone Won't Save Us