in

People are turning iconic photos into art in the style of Studio Ghibli after ChatGPT update


OpenAI just updated ChatGPT’s image generator, and users have begun flooding social media with famous pictures reproduced in the anime style of Studio Ghibli. The results are impressive, but it is not something new to ChatGPT. The AI could do this before. However, the update added several enhancements, including better image handling.

The pictures that people convert cover various topics and media, from movies to news photos to historical events. I’ve sprinkled several examples throughout this article representing some of the best and most recognizable images. As you can see, applying Studio Ghibli’s iconic and whimsical art style gives even darker images, like the JFK assassination photo, a different tone.

It’s unclear how these users are getting ChatGPT to produce these images. When I tried a couple of times, the bot refused me. It said this when I tried to reproduce the album cover for Abbey Road in Ghibli style:

“I was unable to generate the image you requested due to our content policy, which restricts the generation of images based on specific copyrighted content, such as The Beatles’ album cover.” Others have clearly found workarounds for ChatGPT’s copyright filters, but I don’t have time for that.

Before:

After:

Not everybody is pleased about this use of AI, particularly Studio Ghibli Director Hayao Miyazaki. He was the lead on several Ghibli hits, including Spirited Away and Howl’s Moving Castle. In 2016, Miyazaki-san commented on AI-generated animation, saying it “disgusts him.”

“Whoever creates this stuff has no idea what pain is whatsoever. I am utterly disgusted,” he said after watching an example of an animated AI-generated zombie. “If you really want to make creepy stuff, you can go ahead and do it. I would never wish to incorporate this technology into my work at all. I strongly feel that this is an insult to life itself.”

For fair context, he made this comment nearly ten years ago. The example Miyazaki watched was not a finished product and was nowhere close to his animation style. It was a textureless humanoid figure wriggling and crawling on the untextured checkerboard surface representing the ground.

Miyazaki might have a different opinion about the images on the internet that look like well-fleshed-out examples from the studio. They say the sincerest form of flattery is imitation. However, he still might not have a favorable option. Miyazaki may view it as a threat to his work, as many others have, but these creations are a far more favorable representation of AI’s current capabilities than the zombie.

Most of the responses on social media have been positive. Many have commented that the pictures are amazing. People are fascinated by the ability of an AI model to spin a real-life event or image into a capricious piece of anime art. Even more are asking the original posters how to do it. Evidently, I am not the only one having trouble getting the model to cooperate.

Like it or not, be prepared to see a massive flood of these images over the next few days. Well, at least until copyright holders start sending out take-down notices. While most of this seems to fall under fair use, there is little doubt that IP owners see it that way.





Source link

Apple WWDC kicks off June 9 with big software updates

A cleaner future for tires: Scientists pioneer chemical process to repurpose rubber waste