Image Playground in iOS 26 Gets ChatGPT Integration for Anime-Style Images

[MacRumors]

If you've seen the anime-style Studio Ghibli images that people have been making with ChatGPT, you might be interested to know that Apple has integrated this feature into Image Playground in iOS 26, iPadOS 26, and macOS Tahoe.

ChatGPT's anime-style output in Image Playground​

Image Playground doesn't add new art styles designed by Apple, but it does support several ChatGPT Styles. Options include oil painting, watercolor, vector, anime, and print, which complement the existing Apple-created Animation, Illustration, and Sketch styles.

Anime, oil painting, and water color styles in Image Playground​

The ChatGPT options are similar to what you can get if you use ChatGPT 4o for image creation, and the anime style in particular is one that's been popular on social media. In ChatGPT, you can type in any style you want, and Apple has a similar feature with the "Any Style" option. You can select "Any Style" and then add the descriptors that you want, even asking for a realistic image.

This was an image of a cat, and ChatGPT was asked to add a bird and make it realistic.​

You can create images based on text descriptions, or you can upload a photo to have it converted into the style of your choice. When using the ChatGPT option, you will need to give the okay for your iPhone to send the image to ChatGPT.

Vector, print, and any styles (specified as Van Gogh) in Image Playground​

As with standard Image Playground image generation, you can add in additional phrases to tweak the result that you get from ChatGPT. ChatGPT is fairly slow compared to Apple's own on-device image generation capabilities in Image Playground, so expect to wait for the end result.

A photo combined with a text-based request​

Unlike Apple's Image Playground default styles, ChatGPT has no problem with copyrighted content. It will make an image that uses the Studio Ghibli style, or that features Elmo. You're also welcome to include violence, so Elmo can be bloody and hold a knife. Technically, ChatGPT isn't supposed to make violent images featuring children's characters either, but there are simple wording tricks to get around that. Apple does warn before sending any images to ChatGPT, and it is made clear that it's ChatGPT creating the images. ChatGPT has far fewer guardrails than Apple's standard Image Playground tools, which is something to be aware of.

It is worth noting that creating images with ChatGPT uses up tokens, and if you have a free account, you won't be able to create more than one or two. You'll need a paid account to use the feature more frequently.

There is no limit on the number of images that you can generate using the built-in Image Playground options, and it is worth noting that Apple has one other new feature you can use, and that's the "Genmoji" option. You can combine multiple emoji into a single Genmoji, a feature that's new in iOS 26, or describe an emoji character to create something new. Genmoji does not work with photos, and it is a description-only setting. The Genmoji that you create can be added to your emoji keyboard so you can use them in Messages.

iOS 26 is available for developers right now, but Apple plans to release a public beta in July. The software will see a public launch in September alongside new iPhone models.

Article Link: Image Playground in iOS 26 Gets ChatGPT Integration for Anime-Style Images

 

[MacUserFella]

Any weebs here?

 

[klasma]

That doesn’t look particularly anime to me. Ghibli-esque, maybe.

 

[Think|Different]

I tried it and, naturally, the ChatGPT renders were slow as hell.

 

[madebybela]

 

[phenste]

weird to distill the entire genre of anime down to the work of Hayao Miyazaki, especially when he’s spoken about his hatred for AI…whatever

 

[Kottu]

only a limited number of characters can be used for the description while chatGpt mode is used

 

[johnsawyercjs]

In the article, the illustration titled "Vector, print, and any styles" actually contains a repeat of the earlier image titled "Anime, oil painting, and water color styles".

 

[eifelbube]

This was an image of a cat, and ChatGPT was asked to add a bird and make it realistic.​

… and we will not be able to tell with certainty anymore whether the photographic capture of a unique moment is real or just reAIlistic. What have we gained? Yes, i agree, genAI opens new forms of creativity but just to express an idea with a click of a button. If one had learned the tools and techniques in Affinity or Photoshop to be able to place a bird on the head of a cat, one would have also acquired a skill.

 

[7thson]

I haven’t payed much attention to WWDC this year but I would have been pleased to hear that Image Playground was not mentioned and quietly removed, never to be spoken of again. Let a developer pedal this schlock.

 

[jclo]

In the article, the illustration titled "Vector, print, and any styles" actually contains a repeat of the earlier image titled "Anime, oil painting, and water color styles".

Oops, fixed that! Thanks.

 

[Bustycat]

I'm not sure about how limited the "anime" style in Image Playground is, but in ChatGPT the style can be flexible. With some tweaks it can look pretty good and unique, like my current avatar here.

The image model is named "GPT Image 1" and users can select any model for the text input on ChatGPT. I found that o3 and o4-mini-high are better for this.

I tried it and, naturally, the ChatGPT renders were slow as hell.

So that's not different from the speed directly on ChatGPT.

 

[svish]

Good to see these improvements for Image Playground. Expecting it to improve over time.

 

[Porco]

weird to distill the entire genre of anime down to the work of Hayao Miyazaki, especially when he’s spoken about his hatred for AI…whatever

Miyazaki is my favourite filmmaker but I agree. Anime is much more than just Miyazaki, or Ghibli, and even Ghibli has utilised more than just its most famous style. The whole thing smacks of the reductive mechanical risks of AI that whilst immensely powerful also homogenises and lacks some of the random factor that makes things recognisably human.

I’m not opposed to AI tools, but I feel like they are being over-used far too readily and lazily. Hopefully some kind of refinement and equilibrium will occur once it stops being everyone’s new shiny toy and the tech industry move onto something else to be obsessed with.

 

[Haiku_Oezu]

All the Ghiblislop I could ever want, at the touch of my finger…

 

[aloysiusfreeman]

 

[Scotty_Trees]

We built pyramids. We built skyscrapers. We built tunnels under freaking mountains. But this....THIS is the pinnacle of humankind's achievements in all it's glory!!! /s

 

[everlast3434]

I'm going to be using this every single day for sure. Super useful stuff.

 

[cnnyy20p]

That doesn’t look particularly anime to me. Ghibli-esque, maybe.

Unfortunately other anime styles other than Ghibli are extremely otaku/weeb. Ghibli-style is the most universally friendly here. That's why it is so loved.

 

[Deepshade]

Yet another tool that few want or need. What an incredible waste of resources. Why don’t Apple get rid of the ai emoji and pointless art department all together and get some real creatives for design and innovation in.
It would be great to see some concept stuff or new ways of working in the UI.
Or even fixes for every day working issues like a forward delete, or a way to take pictures with just a key trigger word or phrase, that doesn’t involve backflips hacks around the accessibility department.
AI and Siri, by now, should be capable of natural conversation and joined up requests.

This is a total joke!