Say what you want about the company, you have to admit that TikTok’s AI content labeling systems are ahead of any other social media platform. The growing popularity of AI and its equally fast-paced evolution in quality has brought us to the point where artificially created content is everywhere. Sometimes it’s easy to determine when a piece of media is AI-generated but in many cases, it can look like realistic work by an artist or graphic designer. AI-generated content labeling is a strategy that the company adopted back in September 2023, but their filtering systems are receiving an update to point out AI content from other platforms as well. The TikTok CAI-C2PA update is a step in the right direction, and the hope is that other platforms will soon follow suit.
Image: TikTok
TikTok AI Content Labeling Arrives While We Debate the Validity of AI Art
Is AI-content generation a valid evolution of our artistic pursuits and expressions? May have pointed out the inevitably of artificial intelligence leading creative endeavors as the technology behind it gets more and more refined. And yet the majority of any population stays adamant that anything created with AI lacks soul and should never be acknowledged as real creativity. The recent indie movie Late Night With The Devil was a wonderfully novel experience that quickly turned into a must-watch for lovers of horror. However, rumors that AI was used in generating some of the graphics instantly caused many to rate the movie negatively. The bad press has been considerable despite the minimal use of AI, and it does bring to light how the general public truly feels about such technology.
Under such circumstances, being able to see where the blurry lines between “real” art and AI-generated art is something that becomes necessary, so users don’t feel misled into appreciating something that wasn’t human-designed. The TikTok AI content labeling initiative began in earnest last year when the company decided that all AI material generated on the app would be automatically labeled. The company had also updated its policies to require users to add TikTok AI content labels themselves when they reuploaded any material that fell under the umbrella.
Even now, if users choose to keep this information hidden, TikTok retains the right to remove the videos from their platform. As self-serving as we humans are, many refused to adhere to the TikTok AI content labeling system making additional safeguards necessary to enforce.
The TikTok CAI-C2PA Update—How Is The Company Going to Address Content From Other Platforms?
Taking matters into their own hands, the TikTok AI content labels are now going to appear on videos that are generated using other tools like DALL-E. This is where the TikTok CAI-C2PA update comes in. The company has signed up to join the Content Authenticity Initiative (CAI) and the Coalition for Content Provenance and Authenticity (C2PA), and will now add Content Credentials when AI-generated content (AIGC) is detected from other platforms. The C2PA combines Adobe’s Content Authenticity Initiative to track digital media, with Project Origins, which is a Microsoft- and BBC-led initiative that has its eye on disinformation in this new era of digital media.
“The Coalition for Content Provenance and Authenticity (C2PA) addresses the prevalence of misleading information online through the development of technical standards for certifying the source and history (or provenance) of media content. C2PA is a Joint Development Foundation project, formed through an alliance between Adobe, Arm, Intel, Microsoft and Truepic.”
—C2PA
The aim of the joint initiative is to set regulations in place to correctly identify and tag AIGC to curb any spread of misinformation and restrict misleading material that regularly pops up online. The Content Credential technology will provide distinct metadata identifiers on content allowing TikTok’s AI content labeling systems to identify and explicitly tag the material automatically. Any images—and eventually audio and video—with the Content Credential label will be automatically highlighted as AI on TikTok. Content generated with TikTok AI effects will also have a label, in case it is uploaded to other platforms.
TikTok AI Content Tagging Is Not Occurring in Isolation
The TikTok AI content labels may be the first for a social media platform, but it isn’t the only company that has set its sights on these tags. Google and Meta have committed to bringing the technology to their various platforms like YouTube and Facebook respectively. Their adoption strategies have been slower to make an appearance, but both tech giants have openly stated their plans to move forward with such technology. The current emphasis on AI-generated content labeling is being considerably fueled by the upcoming elections in the U.S. and the worry over misleading content being spread across platforms.
Earlier this year, 20 companies announced their joint efforts to prevent artificial intelligence from being a prevalent factor in elections across the globe, according to Reuters. Watermarking and embedding metadata were among the strategies that were put forward by these companies and TikTok was among the list of those who had joined the cause. Politics aside, the threat of AI misuse has been on many minds and different initiatives have been turning up periodically to champion the cause.
The TikTok AI content labeling ambition is a hopeful indicator of how some of the downsides to artificial intelligence can be addressed but it probably isn’t sufficient to fully tackle the problem at hand. Those who want to ignite the flames of chaos always find a way, and it will fall to tech companies to stay on their toes and preempt the moves these miscreants are set to make.