Apple WWDC: New subdomain hints at AI strategy change?

How far-reaching Apple's changes to its AI functions will be is the question on everyone's lips just days before Monday's keynote on June 8. The discovery of a new subdomain on the Apple website and reports of changes in image generation in Image Playground provide initial clues as to what can be expected at the WWDC developer conference.

At first glance, Apple setting up a new address for the browser might not seem that spectacular, were it not for an unusual designation. With genai.apple.com – discovered by bloggers from the news site MacRumors – Apple is using a designation for generative AI that it has so far left to competitors like Google and OpenAI. Assuming the discovered URL is not just a red herring, a possible departure or reorientation of marketing terms could be emerging.

New name on the horizon?

In 2024, Apple wanted to redefine the English abbreviation AI (artificial intelligence) with its term, Apple Intelligence. Given the long wait for features that can compete with those of its rivals and the newly formed cooperation with Google, whose AI models are to be used, Apple's marketing department might want to give the child a new name.

Fittingly, Bloomberg reporter Mark Gurman reports, citing unnamed sources at Apple, that the AI improvements are also expected to affect the image-generating app Image Playground and Genmojis. While the AI-generated mix of emojis and personal ideas and photos is resonating positively, especially with younger audiences, Image Playground was found to have significant design limitations in its initial rollout. To this day, Apple only offers a small selection of image styles, all of which have in common that no realistic images can be produced.

Is Apple loosening the reins?

With the expansion of ChatGPT's image-generating capabilities, the quality of the results has improved. Users have since had the choice of whether to rely on Apple's AI or ChatGPT. Unlike Apple's own models, which generate images locally on the device, ChatGPT uses OpenAI's servers. Nevertheless, even this does not match what users can achieve with the image AI models from Google (Nano Banana) or OpenAI (ChatGPT Images) on the web.

It seems likely that Apple will provide Google's Gemini image models as a result of the cooperation with Google. Gurman's report mentions additional third-party AI models. It is questionable whether Apple will simultaneously loosen the reins on what users can do with the models. So far, the added value of Apple's image AI app has primarily been reduced to the approach of using graphical elements instead of text prompts, which should be welcome, especially for users inexperienced with AI.

Changes to Apple's AI capabilities in its operating systems are primarily expected in voice assistance through Siri. The Shortcuts app is expected to enable AI-powered creation of new actions in the future. AI-assisted functions for accessibility have already been announced.

(mki)