Apple’s Unique Approach to AI
Apple’s marketing lead Greg Joswiak and software head Craig Federighi discuss Apple Intelligence, Siri’s role, and why they believe Apple isn’t competing in the same AI landscape as others.
Apple executives are typically reticent in interviews, often focusing only on their products being the best. However, a candid preview of an upcoming interview offers insights into where Apple positions itself in the broader AI landscape.
The Wall Street Journal’s Joanna Stern recently spoke with Apple’s SVP of Software Engineering, Craig Federighi, and SVP of Worldwide Marketing, Greg Joswiak. A seven-minute segment focusing on AI has been shared ahead of the full interview release. Much of Federighi’s dialogue on Apple Intelligence was similar to that in a previous conversation with iJustine, but Joswiak added new perspectives.
Both executives clarified that Apple has functional versions of contextual Siri powered by app intents, emphasizing that what they showcased was not just a demo. Joswiak even dismissed the notion that they had presented “demo ware,” likely a jab at Daring Fireball‘s John Gruber.
Stern pushed the executives on the practicality of Apple Intelligence and whether they could match the competition, noting that she relied more on competitors’ products.
Here’s Joswiak’s full response:
“Again, it’s important to realize our strategy is a bit different from others. Our vision for Apple Intelligence involves using generative AI as an enabling technology across our operating system. In fact, you might find yourself using Apple Intelligence features without even realizing it, and that’s our goal.
We aim for integration rather than a standalone app called Apple Intelligence. This is distinctly different from a chatbot, a misconception that some have embraced. We chose to provide access through ChatGPT, which we believe is the best option, but our focus is on seamless integration throughout the operating system, resulting in features that I personally use daily.”
His comments reflect much of what I previously posited about Apple’s position in the AI arena. The perception that Apple competes directly with ChatGPT seems misguided.
Federighi adds that Apple need not deliver every technology available. No one questions why Apple isn’t a shopping hub like Amazon or doesn’t compete with YouTube, so it seems peculiar that there’s pressure for Apple to create a chatbot.
You can check out the clip below.
These Aren’t New Arguments
Rather than being a standalone destination, Apple Intelligence serves as an underlying framework that enhances everyday user activities. Users shouldn’t even be aware they’re engaging with AI, let alone with Apple Intelligence.
Image Playground created this image from the prompt ‘WWDC’
The private, on-device Apple Intelligence is gradually being integrated into operating systems beyond its initial feature sets. Developers can utilize this on-device model to achieve results without relying on ChatGPT and compromising user data.
There are whispers of Siri adopting an LLM backend, and with contextual updates via app intents, it appears Apple won’t be launching a standalone AI app or chatbot. Instead, the Apple ecosystem is poised to facilitate personalized, context-aware, proactive interactions across devices and applications.
As previously mentioned, while it may seem that Apple is lagging behind—given Image Playground’s less-than-ideal outputs or Siri’s calendar struggles—I wouldn’t bet against them. As Apple takes its time to ensure its features are primed for launch, others may be rushing to deploy less efficient solutions powered by environmentally taxing resources.
It’s not that Apple is trailing in the AI landscape or even waiting to outpace others; rather, they’re pursuing a distinct path—one focused on delivering robust apps and systems to the iPhone while allowing users access to tools through Visual Intelligence and ChatGPT integration.
For the time being, I will continue using Apple Intelligence daily for my work and look forward to what developers achieve with the Foundation Models framework later this fall.