Amid an escalating debate regarding data privacy and cloud-computing concerns, Apple is leading the charge to shift the narrative about data privacy and computing in the cloud. Sure, from the early years of my Mac to now, a steady stream of things I do online has been transmitted to Apple’s servers, but it has always felt like I, and not Apple, was in control of that exchange. A few short years ago, I gave up the high-maintenance Mac experience in favor of Windows and Android, in part because I preferred the sense of ownership on those open-source platforms – as opposed to feeling like a guest and commodity on Apple’s closed platform. But it appears that Apple might be engineering a way to have the best of both worlds. Now the venerable hardware company is on the verge of launching its latest foray into artificial intelligence – which they’re branding as Apple Intelligence. It’s an interesting dance between skills that utilize cloud AI and keeping your data and privacy niche secure, free from pernicious commodification. How can Apple send your data to its cloud AI servers without losing your ‘apple-pie privacy’ in the process?
Historically, Apple has made a big deal out of its stance that most of the heavy lifting for tasks like photo organisation or enhancement is done right on the device. By comparison, Google or Microsoft tend to favor processing in the cloud. But now Apple has opened the door with Apple Intelligence. Most of the details remain under wraps. The pivot to the cloud brings up inevitable questions about how Apple will play the cloud computing dance without striking a minefield of privacy concerns.
The linchpin of Apple’s new strategy is a new term called ‘Private Cloud Compute’. The term is subtle: note the ‘private’ in the middle which is not usually found in descriptions of regular cloud computing. In a particularly revealing conversation at the WWDC 2024, Apple’s SVP of Software Engineering, Craig Federighi, explained it thus: Apple AI does do some processing locally, but its more complicated things must be sent to our servers. That doesn’t mean we send everything.
And for that, firstly, Jumbo will only upload the data needed to run the computation, which reduces the amount of data you reveal. Secondly, the data that gets shared with the cloud is hidden so that the relevant service cannot identify you – the same IP masking technology that iCloud Private Relay used.
Even better, the data is only stored for a brief time on Apple’s servers. ‘We call that cloud storage,’ Federighi pointed out. ‘It’s also critically important that there’s no permanent storage resident on those servers, and Apple can’t maintain – these servers can’t maintain even logs.’
Apple is setting a new security bar by requiring all its Private Cloud Compute servers to operate using software with publicly published images. This means that the system can be inspected line-by-line by security experts, greatly increasing confidence in its resilience to security problems. Apple devices will only communicate with servers that meet these security standards. If you modify the server software, the local devices will be updated accordingly.
And his remarks emphasize Apple’s continued commitment to privacy, by requiring these stringent security measures even when they make its phones harder to use. No one — not Apple, not anyone else — can access the information used to process your request. It’s also a sign of how much things have changed. Ten years ago, people had to mentally juggle the fact that Google and Amazon were using their devices against them and services that were making their lives easier. It was difficult for them to wrap their heads around the value of collecting one’s data and sharing it with companies who could profit from it. Now the fact that companies like Apple have taken steps to protect users’ privacy is very real, and the efforts they’re making have certainly made a difference.
With Apple venturing into things such as Apple Intelligence, enabling it to join the cloud AI race, it seems that Apple’s careful planning and execution of data processing – achieving unparalleled AI performance without spilling any user secrets – are to become a de facto industry standard. This article was written for Aeon+Psyche, a digital magazine that examines the human condition through the lenses of psychology, philosophy and the arts. It is jointly published with Aeon+Psyche and made possible through the support of a grant from the John Templeton Foundation. The views expressed in this publication can be attributed to the author(s) and do not necessarily represent the opinions of, and should not be attributed to, the Foundation. Funders to Aeon magazine and Aeon+Psyche are not involved in editorial decision-making.
Underneath that is a company that is innovative, but cares about user experience while promoting privacy. And that philosophy is the same in every Apple product, whether it’s the first Apple Intelligence or whether it’s the iPhone that we have today. That’s a philosophy of pushing the limits, but making sure no users are left behind. And we are really trying to change the future. We’re really trying to change technology, and change the future of our users, and change the future of those who want to take their data from them. Effectively, Apple’s work, particularly its efforts in privacy-protecting cloud AI, is a glimpse into the future.
© 2024 UC Technology Inc . All Rights Reserved.