I’m having trouble determining whether the ads I see in my TikTok feed were created by generative AI tools. As someone who spends a lot of time scrutinizing images and videos that are often “said” to be something synthetically generated, some of the promotions I’ve seen definitely raised suspicions. But for several weeks, we hadn’t seen any examples of the AI disclosures required by TikTok’s advertising policy, so there was no way to know for sure.
That’s what irritates me someone You can know for sure if your content was generated by AI. They just don’t tell us anything. And if the companies that claim to support AI labeling efforts actually want AI labeling to succeed, they should probably do something about it.
Let’s take Samsung as an example. After continuing to stream AI-generated videos to my social media channels, I started noticing ads on TikTok that poked fun at the Galaxy S26 Ultra’s privacy display feature. A video that appears to be from the same promotional campaign was published on YouTube, with a folded description revealing that AI tools were used to create the video. By comparison, TikTok’s ads did not indicate whether AI was used. Regular videos on Samsung’s TikTok account (videos that are not actively promoted as advertisements) also lack AI disclosure, despite being labeled as AI-generated on YouTube.
It’s important to note that both Samsung and TikTok are members of the Content Authenticity Initiative, a group that aims to make content authenticity and transparency “scalable and accessible” by accelerating the adoption of C2PA across the industry. In other words, TikTok and Samsung appear to share similar ideals when it comes to labeling AI content. If Samsung intentionally used AI to create the video, it should have notified TikTok at the time the ad was sent. If TikTok had the information, it should have followed the platform’s own advertising policies to ensure users were made aware.
Advertisers on TikTok are only allowed to use content that is “substantially” edited or generated by AI if they make it known. This can be accomplished by applying TikTok’s own AI labels or by adding disclaimers, captions, watermarks, or stickers of the advertiser’s choice, in accordance with the video platform’s business advertising policies.
“When we say ‘substantially modified by AI,’ we mean content that has been modified by AI beyond minor adjustments or enhancements. This includes using real-world images or videos as source material that are significantly modified by AI, such as:
•Content that includes images, video, or audio that is completely generated by AI
•Show the main subject doing something that they are not actually doing (such as dancing).
• Use AI voice clones to make your primary subject say things they didn’t actually say. ”
Samsung did not respond to my request for comment. TikTok pointed me to its AI labeling requirements for advertisers and its C2PA partnership, but declined to provide a statement on the record as to why Samsung’s AI-generated ads were allowed. We do not yet know which stage of this transparency process failed.
I discovered a new development earlier this week. It’s a TikTok ad promoted by UK-based used car retailer Cazoo that we’ve come across before. without it At the bottom of your disclosure, you’ll see the message “Advertiser labeled as AI-generated” next to the “Advertisement” identifier. I already suspected that the ads in question were probably generated by AI, as they all contained strange visual distortions that couldn’t be rationally explained by editorial explanations, such as a dentist’s drill morphing into different shapes or jumping between hands.
It’s been a few days since Samsung’s ad on TikTok was promoted on my feed, so I’m not sure if a similar update has been made. However, AI transparency across Samsung’s TikTok account is generally confusing. Some have TikTok’s own AI labels applied, others include manual disclosures in the fine print of the videos, and some AI-generated examples have no disclosures at all.
Currently, there are no reliable technical solutions to reliably identify AI-generated content, or even human-generated content, at scale. I’ve spent a lot of time thinking about the flaws in authentication standards like C2PA Content Credentials, SynthID, and other provenance-based systems that try to tell users how their content was created. To work effectively, everyone needs to be on board, and that just isn’t happening. This is a problem in this current geopolitical climate when people are struggling to understand what is true and what is not.
But while this is true for online content in general, advertising is a regulated industry that is supposed to play by different rules.
Many of these rules were enacted to protect consumers from misinformation or outright lies by advertisers, such as a law that prohibits cosmetics companies from wearing false eyelashes on models to sell mascara. TikTok beauty influencers like Miqueila Nogueira have learned the hard way that these rules apply when promoting products, and that viewers tend to react badly to dishonest shilling tactics.
While generated videos are not always misleading, concerns about advertising transparency have led the EU, China, and South Korea to introduce AI display requirements in promotional materials. Even companies that do not commit to supporting AI transparency efforts risk future fines if they do not take action.
If large online platforms like TikTok and advertisers like Samsung can’t be honest with each other about their use of AI in such a regulated environment, then anyone can promote as much nonsense as they want. We’re happy to see that at least some ad-specific AI labels are starting to appear on TikTok after reporting ads directly to the companies involved. However, this is a simple two-way system and already needs to be robustly implemented and enforced. without it People like me need to scrutinize every ad in my feed.
