Does Apple want to lower genAI expectations for WWDC?

There’s been a change in tone concerning what to expect from Apple’s forthcoming AI announcements at WWDC, so perhaps it’s time to moderate the hype.

What we’ve been hoping for is an impressive counterattack from the company, one designed to shrug off speculation the company is falling behind on AI. With thousands of engineers and billions of dollars focused on AI research and development, there’s been building expectations of something impressive from the company. However, if Apple industry bellwether Mark Gurman has it right, Apple’s planned announcements, while good, might not quite reach the pinnacle of great.

Is better enough?

That doesn’t mean what’s coming won’t be interesting or noteworthy. Gurman seems to expect some impressive highlights, including tools such as voice memo transcription, summaries of notifications and web pages, and generative AI-powered editing tools.

The latter will apparently work in a similar way to how genAI works in Adobe’s creative apps, which presumably means you’ll be able to generate machine-created images and apply edits using voice/text prompts on your devices. Gurman doesn’t seem to think these will impress regular Adobe users, but given that most of the world’s population don’t use Adobe, it’s reasonable to suppose that for many millions of people, Apple’s tools will be their first exposure to the potential of such technologies.

The gap remains

All the same, despite Apple’s advantages in market reach and platform size, Gurman claims Apple executives still think there is a gap between the current pace of Apple’s genAI development and that of its competitors. He even says this gap is unlikely to close soon, which is perhaps why Apple has been speaking to competitors such as OpenAI, Google, and Baidu.

It’s possible we’ll learn of a deal between OpenAI and Apple at or around WWDC 2024, potentially including integration of ChatGPT natively on the iPhone.  We might also see interesting new features built on the company’s recently introduced tools for accessibility. 

If Apple’s truly not quite there yet, it will not want to disappoint with weak-tea WWDC news — and that makes this a good time to constrain the optimism. This part of the match isn’t over; Apple and AI is a work in progress; and the company’s R&D teams continue to churn out powerful-seeming foundational technologies, including its very own multimodal LLM mode, called Ferret.

The privacy thing

One area of speculation I don’t think Apple hopes to quash revolves around privacy and edge AI. It seems probable that edge intelligence will guide some features, implying that when you do use genAI on your iPhone, the process/data will be kept confidential. 

That is essential if Apple wants to make using these tools a customary part of daily life, particularly in the enterprise space. With that in mind, it is curious that the tone of Gurman’s comments suggest Apple’s focus on privacy and security is limiting what it can achieve with AI — but does at least value the information. Apple is planning its own cloud-based genAI services that should deliver functionality and security, and is investing in highly secure data center processors.

There is a one more card in play that could work in Apple’s favor in the long game: ChatGPT and Google Gemini are server-based solutions, but their future evolution will be constrained by AI regulation and the need to maintain data sovereignty.

These forces will become a barrier to growth, and it remains possible that by focusing on data privacy today, Apple could hold a winning hand by the end of the game. So, while company insiders may be attempting to guide expectation a little lower as we travel toward the new AI iPhone in fall, the game at this table isn’t over yet. Partnership, or even acquisition, could be the next set of cards in Apple’s deck.

Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.