TikTok Launches AI Greensreen, an In-App Text-to-Image AI Generator
TikTok has debuted “AI greenscreen,” a new in-app text-to-image AI generator that allows users to type in a prompt and receive an image that can be used as the background in their videos. The new feature is accessible via the camera screen of the short-form video app.
When compared to state-of-the-art text-to-image models like Google’s Imagen, OpenAI’s DALL-E 2, or Midjourney’s eponymous software, TikTok’s system produces fairly basic results. It only generates abstract and swirling images, which reflect the dreamy nature of TikTok’s suggested prompts such as “astronaut in the ocean” and “flower galaxy.” In comparison, other models can generate both photorealistic imagery and complex and coherent illustrations that appear to have been drawn or painted by a human.
However, the limitations of TikTok’s model may be deliberate. For starters, more advanced models necessitate more computing power, which would be costly and resource-intensive for the company to implement. Second, TikTok has over a billion users and giving all of these people the ability to create photorealistic images of anything they can imagine would almost certainly result in some troubling outcomes.
How TikTok’s AI greenscreen Works
A trial test for the model to generate nudity and gore resulted in two types of output that text-to-image generators frequently attempt to limit. Images based on violent prompts such as “assassination of Boris Johnson” and “assassination of Joe Biden” produce mostly abstract swirls, with a barely recognizable face for the UK’s prime minister (though the man’s familiar blond mop makes caricature especially easy).
Similarly, a nudity request — “naked model on the beach” — yields thematically appropriate colors such as flesh tones, sandy oranges, and ocean blues, but nothing that would make a vicar blush.
What’s notable about the appearance of TikTok’s “AI greenscreen” is how quickly this technology is becoming mainstream. The most recent cycle of development for text-to-image AI began in 2021, with the release of DALL-E by OpenAI. Less than two years later, the technology is already in the hands of millions of people through apps like TikTok.
The filter is currently being used for a few TikTok trends, including one in which you enter your name into the generator to see what your “aesthetic” looks like. Another popular trend is for users to enter their birthday into the generator. Of course, some users are unsuccessfully attempting to determine whether the generator will recreate specific body parts.
This adds to the growing number of new features and modifications by the short-form video platform in its attempt to stay above its competitors in the ever-competitive market.