'We want to get it right': Inside Craig Federighi’s vision for Siri, privacy and Apple’s cautious leap into AI

Apple is about to “go all in” on artificial intelligence, but not in the way you might expect. While rivals like OpenAI and Google push the boundaries of generative AI at a breakneck pace, Apple has opted for a more measured approach. In an interview with WSJ’s Joanna Stern, Apple’s software chief Craig Federighi explained why the company is being cautious as it ramps up its AI capabilities, especially with Siri, its voice assistant.

The Future of Siri: A Smarter, More Personal Assistant

Federighi spoke candidly about Siri’s journey, admitting that while it already processes an astounding 1.5 billion requests per day, there’s much more to come. “This power of Apple intelligence and large language models means that Siri can understand you better than ever before and is going to have access to more tools on your device to do more for you than ever before,” Federighi said. “We looked at this as not how do we build another chatbot and bolt it on the side of our existing experience, but how do we create something that’s deeply integrated and personal,” he added.

However, this transformation won’t happen overnight. Federighi stressed that the improvements to Siri will arrive incrementally. “This is a big lift, and we feel like we want to get it right. This is a many-year, honestly even decades-long arc of this technology playing out, and so we’re going to do it responsibly,” he explained. While some features, such as deeper contextual understanding (like recalling the name of someone you met months ago), are on the horizon, Siri’s growth will unfold gradually over the next few years.

Apple Intelligence: More Than Just Generative AI

The concept of “Apple Intelligence,” introduced earlier this year, goes beyond the flashy chatbot experiences seen elsewhere. It involves a set of advanced AI tools tailored specifically to Apple’s ecosystem, from smarter notifications to more intuitive personalisation across devices. But Apple’s approach, unlike its competitors, is centred on privacy.

Instead of relying heavily on cloud-based AI, Apple is developing its AI to run directly on users’ devices, ensuring data remains private. “We looked at this as not how do we build another chatbot and bolt it on the side of our existing experience, but how do we create something that’s deeply integrated and personal,” Federighi said. This vision relies on the concept of “private cloud compute,” a system where AI models are run locally on your iPhone, with cloud processing only used when necessary, and in a highly encrypted, privacy-respecting manner.

Federighi acknowledged that this approach is more difficult to execute but said it’s essential to protect user data. “For many companies, I think their answer is ‘send us all your data, store all your data with us.’ But as a user, you’re not quite sure, is that all they’re going to do? What are they going to do with it in the future?” Apple, he stressed, wants to ensure its AI models protect user privacy at every step.

Why the Delay? A Deliberate Strategy

Apple’s cautious entry into AI is not due to a lack of ambition but rather a desire to execute each step with precision. With the upcoming iOS 18.1, the first batch of Apple Intelligence features will roll out, but many of the flagship tools shown at Apple’s June event are still in development.

Federighi explained the delay, saying, “We knew at the time when we started out this year that the breadth of our vision for Apple Intelligence… was not all of it landing in September with a bang.” Apple’s philosophy, he said, is to avoid rushing. “You could put something out there and have it be sort of a mess, or… let’s try to get each piece right and release it when it’s ready.”

Privacy-First AI: Apple’s Competitive Edge

In a market where many tech companies are vying to be first, Apple’s emphasis on privacy sets it apart. Federighi was clear that Apple’s AI won’t function like OpenAI’s GPT models, which process every query through cloud servers. Instead, Apple’s AI will function directly on the user’s device whenever possible, tapping into the cloud only when absolutely necessary.

“There are some cases—let’s take a long-form summarisation—an on-device model could try to do that but would probably not provide the best result,” Federighi explained. “If we run that in the cloud, it can both be faster in that case and also provide a higher-quality result. But the important part is that message you asked to be summarised was encrypted to a trusted device in the cloud… and never recorded anything about that data.”

This privacy-first approach, while technically challenging, is core to Apple’s brand, and Federighi took a subtle dig at companies that rely heavily on cloud processing. “It’s hard to build, for one, and I think it also runs counter to many companies’ entire philosophy about how they see data,” he said.

AI Tools: From Funny Notification Summaries to Photo Cleanup

One of the current Apple Intelligence features, notification summaries, has already drawn both praise and humour. Federighi shared an anecdote about the AI summarising a break-up text. “There was a story recently about someone who got broken up with, and it gave a very clinical description of what had happened. No one wants to get broken up with on a text anyway,” he joked.

Another popular feature, photo clean-up, allows users to remove unwanted elements from images, such as a water bottle or microphone. Federighi emphasised that while Apple made this small concession, it remains mindful of maintaining authenticity in photos. “There were a lot of debates internally—do we want to make it easy to remove that water bottle? Because that water bottle was there when you took the photo.”

What’s Next?

While Apple is taking its time with AI, the future looks promising. Siri is getting smarter, privacy is prioritised, and Apple Intelligence is poised to redefine personal computing. Yet, Federighi’s closing remarks were clear: this is a journey that will unfold over years, not months. “This isn’t a one-and-done kind of situation… This is a many-year, honestly decades-long arc of this technology playing out, and so we’re going to do it responsibly.”

For users eager for the next AI breakthrough, it might be a waiting game, but Apple is betting that its slow, careful approach will ultimately pay off in the long run.