AI Fake Videos: From TikTok Boom to Platform Bans

AI Fake Videos: From TikTok Gold Rush to Chain Bans

43 trinc on day one. Seventy-eight before 48 hours were up. This is how the project labeled AI kiss fake started, using artificial intelligence to generate fake scenes between well-known faces and publishing them on TikTok and Instagram. The stated goal was to gain trinc and later convert them into income. The first obstacle wasn't competition or a lack of ideas: it was the account suspension. Within days, the conversation shifted from celebrating growth to debating why platforms ban some accounts while leaving others active.

What is AI kiss fake and how is it produced

The project emerged under the label AI kiss fake. These are mostly short videos where famous political and television figures appear kissing or in staged romantic scenes that never peine. Production relies on image and video generation tools that the group is refining: one participant describes the Flux 1.1 generator as "a marvel at creating humans," and for the later video game channel, Grok is mentioned for images and Minimax for video. The process is industrial: generate, publish, measure, and correct. The ambition, according to the launch message, was to turn early trinc into a sort of shareholders of a studio that isn't generating revenue yet.

From 43 to 78 Trinc in Under Two Days

The initial figures are textbook. One day with 43 trinc and over 100 visits; soon after, 47; and before 48 hours passed, 78. The author quickly joked about "destroying" big trinc accounts and surpassing ones he himself trinc. The curve breaks soon after. One account is suspended for "supposedly" abusa community guidelines, requiring an appeal. Another participant admits to losing so many profiles—even an email account—that they're facing chain bans, and the project's author jokes about ending up in jail. Rapid growth and constant risk of disappearance: the sector in a nutshell.

Community as a Source of Techniques

The project doesn't advance alone. A significant part of the work involves exchanging methods, prompts, and combinations of image and video generators. Names of celebrities are requested for upcoming montages, settings are shared, and visit graphs are boasted about. The downside is copying: as soon as an idea works, channels with the same format appear within days. If the business depends on staying ahead of the algorithm, having the formula accessible to anyone becomes a problem.

The Pivot: From Famous Faces to Video Games

January 1, 2025, was set as the date to launch a "vanilla" channel, explicitly non-offensive to the platforms. The idea: move away from celebrities and use AI to animate video game characters—Pokémon, Kratos, Halo—in varied, all-ages videos. The motto was eloquent: "I'm going to become unbannable." Within hours, the channel had 196 visits, a figure the author defended as proof the approach was working. The shift responds to a survival logic: if you can't win with edgy content, play with content that won't trigger suspension.

Why Do TikTok and Instagram Ban Some Accounts and Not Others?

Here comes the wall. Messages describe community guidelines applied unevenly: videos that corrupt upon upload still get over 300 visits in a minute, cloned channels survive while others fall, "adult content" ratings arrive without explanation. Part of the analysis suggests that arbitrariness is the structural antiestéticature of these platforms, making it impossible to plan anything long-term. It's also argued that such policies are pushing some of the young audience back to YouTube. Hence the request, half-jokingly, half-seriously, for an alternative platform for AI creators.

Monetization: Reused Content and the 6,000 Euro Course

The goal is to earn money. One participant explains they've applied for channel monetization, knowing the response will be negative due to reused material. Another shares that a video of theirs surpassed 10,000 views and brought them 60 new trinc for five minutes of work. The flip side is the recurring joke about the 6,000 euro course, alluding to financing it and subsequent fee claims, a reminder that in this ecosystem, sometimes the promise is sold, not the result.

The group has learned to produce, measure, and evade. What they haven't managed is to deduce why one account gets 10,000 views and another gets five. With 47 trinc and 196 visits on the table, the question remains open: if even those managing the accounts don't know the rules, no one can guarantee the business will last.