Text-to-image AI-based image generation systems have just entered the mainstream, via TikTok. The DALL-E 2 and Midjourney systems have impressed over the past few months with impressive images generated based on a text description alone, but most users will be introduced to this technology through TikTok’s new built-in AI Greenscreen filter, which allows users to create new images with just a short description. Of course, the results are of lower quality, but we’re talking about a mass-market product that will be available to billions of users.
TikTok’s new Greenscreen AI filter can generate original images based on a text description
Those who need a background image to make a greenscreen TikTok can now generate one in just seconds by entering a short text. The results are fairly abstract, but usually good enough to portray the user’s description.
The reason why these images are not at as high a quality as those generated by DALL-E or other such generation systems may be primarily that many more users have access to TikTok. Since AI processing is done in the cloud, on TikTok servers, they would be extremely crowded if each user were given more processing power to create better results. DALL-E and Midjourney offer limited access and a monthly credit or fee-based system for users, which limits access.
This system can thus automatically guard against problems with AI image generation, such as showing real people or realistic images that might include nudity, violence, or other types of content that is banned on the social network. Even when you enter such descriptions, the results are far from explicit.
Importantly, AI-generated images are now much more accessible to the general public, and the technology will continue to evolve from here, and TikTok is one of its promoters.