TikTok will test improvements to its detection systems for accounts posting AI-generated spam on politics and news, financial advice and medical topics. as announced in an editorial article. Testing is expected to begin in the coming weeks and has outlined these categories as areas where misleading content could affect public trust or well-being.
Medical content is one of three designated targets, and it aligns with a Kapwing report I covered in Junewho found that health was among the highest categories of AI-generated content on TikTok, well above categories where creators appear on camera.
TikTok explained that testing is being rolled out at the account level, saying the company is already removing spam on a large scale and has taken down more than 86 million fake accounts in the first three months of this year.
How other platforms reacted
The announcement is part of broader efforts to promote authentic content on social media platforms. Last July, YouTube has updated its monetization guidelines to better meetrepetitive videos, and Meta quickly followed with their own measures against unoriginal content.
Additionally, in June, Google published an article on a system designed to detect and eliminate coordinated groups of accounts posting AI-generated spam.
The headquarters of the C2PA committee
In addition to spam testing, TikTok announced its participation on the steering committee of the Coalition for Content Provenance and Authenticity, the organization responsible for developing the C2PA standard that tracks the origin and changes of digital content.
TikTok mentioned that it was the first video platform to implement C2PA content credentials two years ago. TikTok said the seat on the committee would allow it to help drive adoption of the technology across the industry.
The company announced that it has marked more than 3 billion videos as AI-generated content, using content IDs, creator tags and its proprietary invisible watermark. In May, Google took a similar step by extend SynthID check to search and promising that C2PA verification would be introduced soon.
Why it matters
This is one of the categories that many marketers are focusing on, like healthcare, which already has a lot of AI-generated content. Anyone creating finance, health, or politics-related videos on TikTok is working on topics that TikTok particularly monitors for spam. What this means in practice will depend on how TikTok conducts the test, and the details are not yet clear from the announcement.
Looking to the future
The details to watch out for are how TikTok defines an eligible account and what follows a detection, neither of which the announcement specifies. There is also no firm start date beyond the coming weeks. On the C2PA side, TikTok has not announced any changes to its existing labeling system.
Featured Image: Mamun_Cheikh/Shutterstock
