PARIS: The latest internet obsession is one impossible to escape on social media: images of well-known figures in plastic toy packaging alongside evocative props… welcome to the meme of AI action figures.
The figurines aren’t real, but thanks to ChatGPT’s new image generator they look genuine, and they are flooding platforms from TikTok to LinkedIn.
There, you’ll see a “toy” image of football star Lionel Messi alongside a ball and trophies, or a caricature of US President Donald Trump alongside a MAGA cap and a sign reading tariffs.
Some celebrities have done versions of themselves. For instance, actor Brooke Shields posted a doll version of herself with a miniature dog and needlepoint kit to her 2.5 million Instagram followers.
Many ordinary users have done likewise, posting packaged images of them as figurines — after inputting photos of themselves into ChatGPT’s image generator.
While pervasive, and in most cases fun, the trend raises questions about copyright, and potential risks in handing biometric data over to a generative-AI company.
ChatGPT does the rendering for free, but requires users to sign up to the imaging platform used, adding to OpenAI’s renown as leader in consumer uses of artificial intelligence — and also for sucking up massive amounts of user data.
Anais Loubere, an expert on social media and founder of the agency Digital Pipelettes, said the boom was riding on a “technological breakthrough” in AI image rendering.
Before the action figure meme, AI artwork inspired by famed Japanese animation outfit Studio Ghibli — maker of Oscar-winning Spirited Away — flooded the internet at the end of March, posted by users subscribed to OpenAI’s paid service.
Ahlem Abidi-Barthe, an online marketing professor, told AFP that the key to the success of the trends was that “they can be in personalised formats”.
“That tickles the egos of consumers,” she said.
The Ghibli and action-figure memes also tapped into childhood nostalgia, “which contributes to extreme virality”.