The most valuable company in the world has just showcased the result of months of research and vast quantities of modern computing power. And it’s a kitsch AI image of someone’s pet dog with a bunch of mistakes in it.
Yep, our first glimpse of output from Apple‘s AI image generator, Image Playground, is more than a little underwhelming. It seems the makers of the best AI image generators won’t have too much to worry about. But that might be the idea. Because Apple could do it better if it wanted to…. right?
Apple is a little bit late to the AI image generation game. It’s been over two years since Midjourney and DALL-E 2 launched and over a year since Adobe Firefly appeared on the scene. Access to AI image generation is now available everywhere from Microsoft‘s Bing search engine and Copilot to Meta social media apps. Apple’s finally joining in with plans to launch its text-based Image Playground for iOS.
Bloomberg’s Mark Gurman says the Apple AI image generator will be rolled out in iOS 18.2, probably in December, along with Genmoji for generating custom emojis. So far, all we’ve seen from it is theone image above, which, according to Wired, was made by senior VP of software engineering Craig Federighi for his wife to celebrate their pet dog Bailey’s birthday.
The picture of a dog wearing a party hat is cartoon-like, looking quite like DALL-E 2. The image also has a watermark, which Apple has said it will always include when it publishes AI-generated images to identify it as such, not that we really need it in this case. The image has that cheesy soft AI glow to it, part of the dog’s collar is missing, and the candles just look weird.
Apple’s AI image looks almost endearingly cartoonlike compared to what image generators like Flux are capable of (Image credit: Secret_Ad8613 via Reddit / AI-generated)
People have been quick to poke fun and suggest that the world’s biggest tech company has ended up at the bottom of the pile when it comes to AI image generation. Over on Reddit, the image was described as “the Finewoven of Apple intelligence features” and even as “Metaverse Zuckerberg with Eiffel Tower bad”.
But I suspect that Apple may be going for a non-realistic look for its AI image generator on purpose. The Cupertino giant has been cautious about AI. And by taking its time, it’s avoided the controversies that have blighted Google, Meta and X’s rushed and messy image generator launches, which saw the internet flooded with gun-toting Marios and much worse.
Daily design news, reviews, how-tos and more, as picked by the editors.
The Google Gemini AI image generator launch had some issues (Image credit: Future)
Now that it’s bringing the tech to iPhone, even the name suggests that Apple intends Image Playground to be a fun and light-hearted feature, perhaps generally aimed at younger users. It’s not exactly out to compete with Midjourney, Flux or Adobe.
Pictures generated by Image Playground include EXIF metadata to identify them as made with AI, as will Clean Up, the new object remover tool for photos. Apple’s made much of Apple Intelligence features working on device rather than in the cloud, but it’s not clear if that’s the case with Image Playground.
At the Apple iPhone 16 launch event on Monday, Apple revealed plans to launch the first Apple Intelligence features with iOS 18.1, expected in October. These will include writing tools for generating and summarising text, suggested replies in Messages and a tool to record and transcribe phone calls.
