One AI-generated human race
None of these people exist, but that hasn’t stopped them from posting profusely on TikTok
In recent days and weeks, unusually similar AI-generated videos depicting nonexistent Black people accompanied by captions such as “dear white people” and “hit follow if you want peace” have proliferated on TikTok. These videos are invariably accompanied by the same set of hashtags, and posted by accounts that appear to post no content other than the repetitive AI-generated videos in question and occasional random spam. Most of the videos rack up a few thousand views, although a handful have garnered substantially higher engagement and been more widely recommended to users.
Over the period from November 11th to November 16th, 2025, at least 45 TikTok accounts posted one or more AI-generated #OneHumanRace videos. 19 of the 45 accounts have handles that are variations on the name “Holly”, frequently with additional O’s. Multiple accounts named Clara, Doris, and Yuki also turn up, the latter with variable numbers of U’s. Some, but not all, of the accounts deleted the AI-generated #OneHumanRace videos between the time they were first observed and November 16th, 2025.
The video above, featuring an AI-generated outdoor scene of a group of younger Black people and the audio “I don’t know which of y’all white people need to hear this today, but I love y’all; if you don’t have a problem with my skin, please hit that follow, we need to be friends” is representative of the content posted by the spam network. The majority of the spam accounts have posted dozens of such videos, with slight variations on the same theme and message, generally ending with a suggestion to follow the spam account. Some of the videos feature indoor settings such as kitchens, grocery stores, and libraries as opposed to street scenes.
The majority of the accounts in the network use still images from the AI-generated videos as profile images, sometimes with added pixellation or blurring reminiscent of an analog television. A few of the accounts use other AI-generated face shots, plagiarized photographs, or drawings as avatars instead. In the case of at least one account, @hooolly90, the profile image has been swapped from an AI-generated image to a graphic of a bottle, accompanied by the deletion of the AI-generated videos formerly hosted by the account.
Each of the AI-generated videos posted by this network is accompanied by the same three hashtags: #OneHumanRace, #RespectAll, and #LetLoveLead. Of these three tags, only #LetLoveLead appears to have a significant amount of authentic content associated with it; the other two hashtags show up almost exclusively alongside the spam network’s synthetic videos. The more popular hashtag is presumably included in order to make the content more algorithm-friendly and increase the odds of the videos being seen by genuine TikTok users.
At least twelve of the accounts in the network have since moved on from posting AI-generated videos of nonexistent Black people to other forms of repetitive content. Much of the new material is obvious commercial spam hawking various forms of merchandise, often Christmas-themed. Other types of new content turn up as well, including at least two accounts that have started posting extremely similar videos about tinnitus in Spanish.
A handful of the accounts in the network also have older repetitive posts unrelated to the AI-generated video spam. The older spam consists largely of videos of landscapes and other outdoor scenes with the caption “so you are living your big dream and this song plays in your mind”. Some of the accounts also have early travel-related videos, depicting scenes from airplanes, hotels, and resorts.
As mentioned earlier, the AI-generated videos posted by the network are extremely repetitive, with variations on the same handful of scenes appearing over and over: small groups of people walking down streets, women cooking, and women walking past bookshelves in a library are recurring themes. The videos in each category are basically all the same, with minor variations in wording, clothing, and background imagery, such as the library scenes in the collage above. This suggests that the operators of the spam network used the same prompts over and over to generate these videos, rather than crafting individual prompts for each.
Many of the network’s videos contain obvious indicators of their artificial origin. In some cases, watermarks from OpenAI’s Sora are still present on the videos, confirming not only that the videos are AI-generated but identifying the specific AI model used to produce them. Other videos, while lacking watermarks, include artifacts that demonstrate their inauthenticity, such as shopping carts that melt together, and grocery stores with nonsensically numbered aisles.
Although the origins and goals of this spam network remain unclear, the repetitive nature of the content and the pivots between different types of spam suggest that the #OneHumanRace videos are a form of engagement farming rather than sincere messaging. The fact that the most recent spam from many of the accounts involves blatant product promotion underscores this, and indicates that, rather than being sincere, the #OneHumanRace videos are simply an effort to build an audience for the accounts for future advertising purposes.
OOooooooo!!! Hope your step-away was helpful, and welcome back!