Text-to-image Right now, AI systems are exploding in both capability and appeal, and what greater evidence than their inclusion in the world’s hottest app: TikTok!!
The video platform recently introduced a new “AI greenscreen” effect that enables users to put in a text prompt that the program would subsequently produce as an image. This picture may then be used as the backdrop for a movie, which might be a very valuable tool for makers.
When compared to state-of-the-art text-to-image models like Google’s Imagen, OpenAI’s DALL-E 2, or Midjourney’s namesake program, TikTok’s method produces quite rudimentary results. It only generates abstract and whirling visuals, which reflects the dreamlike character of TikTok’s recommended suggestions such as “astronaut in the water” and “flower galaxy.” In contrast, some models can generate both photorealistic pictures and intricate and cohesive drawings that seem to have been created or painted by people.
However, the restrictions of TikTok’s paradigm may be deliberate. For starters, more complex models need more computer power, which would be costly and resource-intensive for the organization to adopt. Second, TikTok has over a billion members and giving all of these people the ability to generate lifelike representations of whatever they can conceive will almost surely result in some terrible outcomes.
What’s striking about the advent of TikTok’s “AI greenscreen” is how quickly this technology is becoming widespread. The most recent cycle of development for text-to-image AI started in 2021, with the release of DALL-E by OpenAI. Less than two years later, the technology is now in the hands of millions of people via apps like TikTok.
Given the potential for both good and evil in these systems, things are just going to grow weirder from here on out.