Apple's Major Debut At WWDC 2024: Apple Intelligence

Apple has finally ventured into the artificial intelligence and machine learning space with the unveiling of Apple Intelligence at its Worldwide Developers Conference 2024.

While the tech giant did reveal significant changes, updates, and new versions of its MacOS, iOS, VisionOS, and iPadOS, Apple Intelligence is what looked most interesting, given the approach the company has chosen.

"It draws on your personal context," said Senior Vice President of Software Engineering, Craig Federighi of Apple Intelligence, during the WWDC 2024 keynote address.

It's a never-before-seen spin on AI and machine learning, pointing it inward at users and their device ecosystems.

Apple Intelligence works by using all of your user data on the iPhone, Mac, or iPad to carry out tasks on your behalf, saving you time. They've trained their large language models to look out for contextual information in the same way we do while speaking to others and recalling, referencing, and cross-referencing our own knowledge.

Apple promises to build AI around its users and their ecosystem, eliminating the need to switch between apps. Apple Intelligence will be going through all the data on your phone and creating context appropriate responses to help you more easily navigate the real world. So, not only will the platform look through everything on your phone, but also what's on your screen.

The company's pitch is more about convenience and less about innovation.

Putting AI on a phone is no easy feat, of course, and that's where Apple says its 'M' family of chips along with the A17 Pro, shine. They're the "computational foundation to power Apple Intelligence,” said Federighi.

These chips will be able to provide the power necessary to deploy AI across the Apple ecosystem, according to the company. The systems built into the Apple products comprise large language and diffusion models specialised for everyday tasks that can "adapt on the fly" to what the user is doing at a given moment.

Apple says that its AI will use on-device semantic indexing. It's a system that finds the information (in this case, your user data) for a query, but it does so by matching the intent and meaning behind it. Apple Intelligence then feeds it to the company's on-device generative models, which will return answers that the company says will best assist the user.

Apple says that for larger, more complicated tasks, Apple Intelligence will contact LLMs based on the company's own cloud servers to better assist users. Of course, that means user data being kicked around different servers.

The company claims that users within the Apple ecosystem have nothing to fear. Apple has long claimed that data created by users is theirs alone, and they have the ability to choose who they share it with. The company claims that it is now extending those protections to the cloud, in what it calls Private Cloud Compute.

"Private Cloud Compute allows Apple Intelligence to flex and scale its computational capacity to draw on larger, server-based models for more complex requests," said Federighi.

Essentially, if a query is too complex for on-device processing, Apple Intelligence will forward the request to the company's large language models based on their own exclusive servers to retrieve answers. Federighi stated that Apple will only forward data relevant to the query to the servers, without storing it.

The company has also said that Private Cloud Compute will only be used to fulfil requests made by users. Apple has also made it possible for independent experts to examine the code that runs on the servers for them to verify privacy.

"Private Cloud Compute cryptographically ensures your iPhone, iPad, and Mac will refuse to talk to a server unless its software has been publicly logged for inspection," Federighi said. This would essentially mean that all of Apple's user queries are routed through their own servers, which use Private Cloud Compute.

This approach, the company claims, is setting a new standard for privacy in AI.

Apple's focus on contextualising your data and making AI work around you within your ecosystem is where the company's real advantage becomes apparent.

While the $3 trillion company paved the way for AI in our operating systems with the launch of Siri in 2011, it quickly fell to the wayside thanks to Google's Assistant and Amazon's Alexa. Apple Intelligence might be the company's chance to refocus the spotlight on itself.