Apple Intelligence: Apple's AI plan, on-device + Private Cloud Compute
In one sentence At WWDC Apple unveils Apple Intelligence: on-device models on A17 Pro/M-series devices, fallback to verifiable 'Private Cloud Compute', ChatGPT integration for hard queries.
Apple arrived late to the generative AI race, and at WWDC 2024 tries to catch up at once. It announces "Apple Intelligence", a mix of AI models integrated into iOS, iPadOS and macOS: writing assistance, email summaries, image generation (Genmoji), a finally-smart Siri.
The approach differs from OpenAI or Google. Apple uses small models running on your iPhone (requires A17 Pro or M1+). For hard requests, it sends data to special Apple servers (Private Cloud Compute) that claim to never see your data in clear, with publicly verifiable code. For even harder tasks, it offers to hand off to ChatGPT, but only with your explicit consent.
Practical result: an AI assistant deeply integrated into apps (Mail, Notes, Photos, Messages), with privacy as a marketing focus. The rollout is gradual (from iOS 18.1, October 2024), initially US English only, Italian arriving in 2025.
Companies
Apple, OpenAI
Tools
Apple Intelligence, Private Cloud Compute, Siri
Tags
Sources