mice-contact

For my sins, I do occasionally scroll through TikTok. In between the usual suspects of ear worm music loops and memetic dances I generally lack the co-ordination to recreate but not the determination, there is a spattering of AI-generated content. On my 'For You' page, this usually takes the form of image filter videos that twist users' selfies into vaguely resembling a frame from a favourite anime. Well, ChatGPT's recent image generation update can now spit out pictures that look like it's traced Studio Ghibli's homework.

The update specifically brings image generation to , refining a number of things AI image generators have historically struggled with, , or even [[link]] . The update also allows users to refine image results through the chat interface, and the AI is apparently now better able to generate consistent variations on a theme. For example, one of OpenAI's demo videos shows GPT-4o generate a penguin mage in various styles, including a low-poly look, a reflective metallic get-up, and looking like a wargaming miniature.

Premium users are already able to get their hands on GPT-4o's style-consistent capabilities, . Free users on the other hand [[link]] will have to wait; after the update's release earlier this week, to explain the delayed roll-out to all tiers of users, writing, "Images in ChatGPT are wayyyy more popular than we expected (and we had pretty high expectations)" (via ).

Judging by my flooded social media feeds, AI-generated images in the style of Studio Ghibli's animated films is the clear viral favourite. This is despite the copyright filters apparently rolled out as part of this update. report that ChatGPT would not generate a Ghibli-fied rendition of The Beatles Abbey Road album cover, instead displaying the following note in response to their prompt: "I was unable to generate the image you requested due to our content policy, which restricts the generation of images based on specific copyrighted [[link]] content, such as The Beatles' album cover."

Many users have , and even OpenAI CEO Sam Altman recently changed his . As to why GPT-4o can so consistently generate images in a number of recognisable styles, that the model was trained on "publicly available data" alongside using data it already has access to as a result of the company's partnership with various companies like Shutterstock (via ).

OpenAI’s chief operating officer, Brad Lightcap, told the Wall Street Journal, "We’re [respectful] of the artists’ rights in terms of how we do the output, and we have policies in place that prevent us from generating images that directly mimic any living artists' work."

Legendary animation director Hayao Miyazaki is still very much alive, and famously took a dim view of early applications of AI. , Miyazaki was shown a demonstration of a rudimentary 3D zombie model animated using AI by developers that also say they hope to one day create "a machine that can draw pictures like humans do." Many retellings of this moment focus on Miyazaki saying, "I would never wish to incorporate this technology into my work at all. I strongly feel that this is an insult to life itself." However, the context often missing is that these words are preceded by Miyazaki talking about a friend with limited mobility, making allusions to the horror genre's often insensitive depictions of disability.

I don't want to put words in Hayao Miyazaki's mouth, so I'll speak for myself here. While GPT-4o's image generation capabilities are an impressive novelty, it also turns my stomach. I'm personally friends with a number of professional artists and I fear this update is simply going to embolden the very worst of their clients. I hope I'm wrong, but cheapskate companies may feel like they can continue to devalue creative skillsets, and I worry that we're all going to be caught in the resulting AI-slop landslide.