Apple Unveils AI-Powered Image Playground Feature

Apple shares first example generated image made by Apple Intelligence: Craig Federighi's dog - 9to5Mac

Apart from some marketing materials, we have not yet seen Apple Intelligence image generation features in action. Image Playgrounds and Genmoji are not currently available in the iOS 18.1 Apple Intelligence betas; they are rumored to launch as part of the iOS 18.2 cycle later this year. However, Apple has today shared a real-world example with WIRED magazine. Pictured above, adorned with a AI watermark applied by WIRED, it depicts Apple SVP Craig Federighi's dog. To be clear, Apple sent the un-watermarked image to WIRED to be published; WIRED applied the 'AI-Generated Image' disclaimer text across the center of it, in accordance with their website's policy on featuring AI works. When Image Playgrounds is available to customers, there will be no visual watermark applied to the photos. EXIF metadata included in the file will signal it was made with AI, as we've already seen happen with the Photos' Clean Up feature. This means iOS and other apps will be able to indicate which images are generated by the AI tools. However, file metadata can be removed using commonly available tools, and similarly if the image is screenshotted, any metadata of the original image is lost. In terms of risk of deception, though, Apple's image generation suite will not fall foul of being mistaken for reality. This is because Apple is intentionally starting with a limited set of non-photorealistic styles: users can pick from "animation", "sketch", or "illustration" style. Craig's dog image appears to have been generated using the animation style, which are seemingly meant to look like Pixar-esque CGI models. It's hard to judge how much the AI creation actually resembles Craig's dog, named Bailey, as we don't have a photo of her to compare against. In the AI image, Bailey is situated in front of a cake and wearing a party hat, as Federighi made the image for his wife to celebrate Bailey's recent birthday. Image Playgrounds lets user enter a text prompt to customize the kind of output they want. When it ships, Image Playgrounds will be available as a standalone app, as well as built-in to communication apps like Messages; concepts relating to the current conversation context will automatically be suggested. You'll need an iPhone 15 Pro or later to get Apple Intelligence, or an iPad or Mac with the M1 chip or later. The first round of Apple Intelligence features will launch alongside iOS 18.1 in October, in the United States. Other variants of English will be supported starting in December, with other languages coming onboard next year. Apple Intelligence features will not be available to customers in the European Union, until Apple can resolve compliance with EU laws. In an interview with WIRED, Federighi said "we do have to comply with regulations, and there is uncertainty in certain environments we're trying to sort out so we can bring these features to our customers as soon as possible".

Apple Shared Its First Public AI-Generated Image. It's Craig Federighi's Dog

Apple's forthcoming iOS 18 is loaded with a slew of artificial-intelligence-powered capabilities. The new mobile OS can rewrite your text messages, summarize emails, and identify objects in photos. But one of the most fun features is Image Playground, which generates cartoon-like illustrations based on a text prompt. While Apple has shown examples of its output during its keynotes, demos, and product videos, we have yet to see a real-world example of an Image Playground character until now. Apple has shared with WIRED the first-ever example created by Image Playground that it's shown outside of its pre-recorded keynotes and marketing materials. This image of an adorable small dog smiling in front of a birthday cake wearing a party hat isn't just any random pup. It belongs to Craig Federighi, senior vice president of software engineering at Apple, who created it for his wife in honor of their pet's recent birthday. Federighi referenced the illustration during his interview with WIRED's Lily Hay Newman about Apple's Private Cloud Compute, the secure server environment the company built to handle AI task requests that cannot be processed on a consumer's device. Apple reps shared it later. It's WIRED policy to clearly identify any AI-generated images we publish, so that's why you see the watermark on the image. Image Playground is debuting at a time when generative AI tools have been creeping into software from all the major tech companies, as Microsoft, Google, and Meta have released AI-powered software that focuses on productivity and creativity. While Apple's approach in iOS 18 also prioritizes the practical side of AI, the company has included some purely fun apps too -- Image Playground being a prime example. It exists as a standalone app, but you can also access it through Messages. To generate an image, you can either type a description of what you want to see, choose a photo of someone from your photo library, or pick from some preloaded concepts. You can also opt between three different styles including Illustration, Sketch, and Animation. The feature is not to be confused with Genmoji, which allows you to generate custom emoji directly from the keyboard using text prompts. Since neither of these generative AI features are available in beta just yet (although a few other Apple Intelligence features are with the iOS 18.1 developer beta), the only examples we've seen of Image Playground's and Gemoji's output have been tightly controlled by Apple. Until the features are released, Federighi's adorable dog is the closest we'll get to seeing an example that exists beyond heavily produced and edited marketing materials. And I'd say it's a pretty good one, and certainly not as creepy as the examples we saw during the keynote at WWDC. Good job, Craig.

Apple Shares First Example of Image Playground in Action, and It's Based on Craig Federighi's Dog

Apple has shared its first "real-world example" of Image Playground, the upcoming Apple Intelligence feature that generates cartoon-like illustrations based on a text prompt. The picture was apparently made by Apple's senior VP of software engineering Craig Federighi for his wife, in honor of his dog Bailey's recent birthday. The picture shows a cute dog wearing a party hat and smiling behind a birthday cake. Apple shared the picture with Wired, which includes a watermark on all AI-generated images that appear in its publications as a matter of policy. Pictures made with Image Playground include EXIF metadata in the file that also indicates it was made with AI, similar to images edited with Apple's new Clean Up tool in the Photos app in iOS 18.1. Apple describes Image Playground as a "fun" feature that can produce original images in seconds right within apps. Users can create an entirely new image based on a description, suggested concepts, or a person from their Photos library. From there, users can adjust the style and make changes to match a Messages thread, a Freeform board, or a slide in Keynote, for example. According to Bloomberg's Mark Gurman, Apple plans to make the Image Playground feature for generating images and the Genmoji feature for generating custom emoji available in iOS 18.2, which will likely be released in December. Apple plans to introduce the first Apple Intelligence features with the release of iOS 18.1, which is expected to launch in October. Among the new capabilities are writing tools for generating and summarizing text, as well as a feature that provides concise summaries of notifications. The Messages app will gain the ability to suggest replies, while a new function will allow users to record and transcribe phone calls. Lastly, the Photos app will benefit from the aforementioned "Clean Up" tool, which is designed to swiftly remove unwanted objects from images.