It feels like just yesterday we were marveling at AI's ability to conjure up images and text from thin air, and now, it's becoming a standard feature on platforms like TikTok. But as AI-generated content floods our feeds, a crucial question arises: how do we know what's real and what's a digital creation? TikTok is stepping up to address this, especially as we move through 2024, with a clearer policy on labeling AI-generated content.
Back in March, TikTok updated its content policies, making it a requirement for users to disclose deepfakes and AI-generated content. This could be done through video captions or by using a special sticker. The goal, as shared by social media consultant Matt Navarra, is to help "prevent content from being removed." This suggests that transparency is key to keeping content on the platform.
More recently, TikTok has introduced a new toggle switch right before you upload a video, nestled under the 'More options' section. This simple switch, labeled "AI-generated content," is a direct way for creators to flag their work. When you flip this switch, a pop-up appears, reminding creators that they need to label content that depicts "realistic scenes" and was made with AI. And, as a stern but necessary warning, it reiterates that incorrect labeling could lead to content removal. It's a proactive step to ensure viewers aren't misled.
This isn't just a TikTok-specific initiative, though. It's part of a broader tech industry push. TikTok is now planning to label images and videos generated by AI from external sources, not just those created using its own in-app tools. This is being achieved through a digital watermark technology known as Content Credentials. Think of it as a digital fingerprint for AI-generated media.
This move is particularly significant given concerns about AI-generated content being used to influence elections. TikTok was already part of a group of 20 tech companies that pledged to combat this issue. By adopting Content Credentials, a technology from the Content Authenticity Initiative (C2PA), TikTok can attach metadata to AI content, making it easier to identify and label.
So, what does this mean for us as viewers and creators? For creators, it's about being upfront. Whether you use Adobe Firefly, TikTok's own AI image generator, or tools like OpenAI's DALL-E, if your content is AI-generated and depicts realistic scenes, you'll need to disclose it. The platform is even looking to extend this to pure audio content soon.
For viewers, this means a more transparent experience. You'll start seeing a label under the creator's username that clearly states "AI-generated." This labeling is being rolled out in phases. Initially, it will focus on content that already includes Content Credentials. As more platforms adopt these credentials, AI identification will become more widespread across social networks.
It's also worth noting that TikTok is setting boundaries. They've made it clear that users can't profit from AI-generated images of well-known individuals, and any "synthetic" content must be explicitly labeled as "synthetic," "fake," "unreal," or "modified." This is a response to the growing realization that AI can blur the lines between fact and fiction, posing societal and personal risks.
This push for transparency aligns with global regulatory trends. The European Union, for instance, is pressuring social media platforms to label AI-generated content to combat misinformation. While some platforms have voluntarily signed accords, others, like Twitter, have faced warnings about potential sanctions if they don't comply with new digital content laws.
Ultimately, TikTok's 2024 AI disclosure requirements are about fostering trust. By making it easier for creators to label their AI-generated content and by implementing automatic labeling for content with Content Credentials, TikTok is aiming to create a more honest and understandable digital space for everyone.
