Apple announces Apple Intelligence at WWDC 2024

Background

On the same day, Apple described its models: an on-device model of about 3 billion parameters and a larger server model.^[2]

What happened

On 10 June 2024, at WWDC, Apple announced Apple Intelligence for iOS 18, iPadOS 18 and macOS Sequoia.^[1] The features included Writing Tools, Image Playground, Genmoji, natural-language photo search and a Siri that can understand onscreen content and act across apps.^[1]

Key details

Private Cloud Compute was introduced to run larger models on Apple-silicon servers. Devices connect only to server software that is publicly logged.^[1]

Apple also announced a ChatGPT integration powered by GPT-4o, which is free to use without an account.^[1]

Apple said the features would arrive in beta that autumn, in US English, on iPhone 15 Pro and on M1-or-later iPads and Macs.^[1] CEO Tim Cook said: "Apple Intelligence will transform what users can do with our products."^[1]

Apple's approach separates work by where it runs. Simple requests can be handled by a model on the device itself, so the work stays on the device. Larger requests need more computing power than a phone or laptop provides, so they go to server models. The design depends on the servers not keeping user data and on their software being publicly logged, so that outside experts can check it.^[1]

Writing Tools rely on language models that can read a passage and then rewrite or shorten it. A notification summary works in the same way, condensing several messages into a short line.

When the features were announced they were described for US English only.^[1] Apple also said they would run on the iPhone 15 Pro and on M1-or-later iPads and Macs, so older devices were not part of the plan.^[1]

Impact

The announcement described three parts of the system: models that run on the device, a larger server model in Private Cloud Compute, and an optional ChatGPT integration.^[1][2]