TikTok to Automatically Identify and Label Increased AI-Generated Content in Its App

TikTok is enhancing its approach to automatically labeling AI-generated content within its platform, even when created using third-party tools. The company plans to implement content credentials, a digital watermark that identifies the use of generative AI.

Currently, TikTok requires creators to disclose "realistic" AI-generated content, but enforcing this policy can be challenging, particularly when creators utilize external AI tools. With the rise of content credentials across the AI industry, TikTok's automated labeling system aims to bridge these enforcement gaps.

Content credentials, often referred to as a “nutrition label for digital content,” embed "tamper-evident metadata" that can trace the origins of images and the AI tools used to create or modify them. This metadata allows users to view the history of AI-generated content encountered on compatible platforms.

TikTok claims it will be the first video platform to adopt content credentials, although widespread use may take time, as many companies—such as Google, Microsoft, OpenAI, and Adobe—are just beginning to integrate the technology. Meta has also announced plans to utilize these standards for labeling on its platform.

However, it is important to note that systems relying on metadata, including content credentials, have limitations. OpenAI has indicated that this technology “is not a silver bullet,” as metadata can be inadvertently or intentionally removed. Additionally, labels may go unnoticed if users don't read them. To combat this, TikTok has partnered with the fact-checking organization MediaWise and human rights organization Witness to initiate media literacy campaigns. These initiatives aim to educate TikTok users about the significance of labels and the potential for misleading AI-generated content.