OpenAI's viral "Ghibli" image generator is now available to everyone
OpenAI's new image generator, already in the news thanks to its approximation of Studio Ghibli art, is now available for all users on the platform.
While the company hasn't officially announced it, an older post on X by CEO Sam Altman suggested that free users will have a limit of three image generations per day.
Launched in March, the image generator is responsible for the latest "Ghibli" trend on social media, where users share "Ghibli-fied" portraits of themselves or others because why not?
The trend went so viral that Altman posted that their "GPUs are melting," because of the number of generation requests.
It has sparked an interesting discussion online about copyrights and AI liabilities when it comes to infringing said copyrights.
Evan Brown, an intellectual property lawyer, says the generator operates in "a legal grey area," in an interview with TechCrunch.
You can't sue for someone appropriating style but can build a case against using copyrighted work to train image models. This topic is highly debated in several courts - whether using copyrighted works for training models falls within fair use.
Some privacy pundits have also suggested this could be OpenAI's way of gathering high-quality image data. Though this is still a conspiracy story, for now. If you intend to try the trend yourself, don't share any personal information or photos.
Several Reddit users have reported seeing an error message saying GPT could not generate the images "due to copyright and intellectual property concerns," but these seem to be inconsistent, with several more still posting Ghibli images on social media.
Source(s)
