Rumours regarding Apple’s new AI strategy were confirmed at its annual World Wide Developer Conference (WWDC) this week. On June 10, the keynote speech announced a suite of new AI features, with a focus on stringent privacy measures.
Apple Intelligence
The tech giant has branded the new features “Apple Intelligence” (cleverly abbreviated to AI), to be available on devices later this year. These features will be compatible with the A17 Pro chip (found on the iPhone 15 Pro and M1 and later Apple Silicon Macs, for example).
AI features will include:
- Personalised prioritisation of notifications on iPhone
- Generative writing capabilities across various apps
- Image generation
Apple also claims its generative AI is tailored to individual contexts, unlike AI competitors, by leveraging a combination of on-device and server-powered models. For example, it can identify related contacts, provide directions, and locate relevant files when reading an email.
Privacy Commitment
A key part of Apple’s AI announcement is its stance on privacy. It claims that, even when providing personalised responses, it will maintain user privacy – either by keeping data on-device or using a no-knowledge server approach.
Apple plans to launch new server infrastructure to support more advanced features, which will be known as Private Cloud Compute (PCC). This will allow complex generative AI models to process user data without compromising privacy. Apple states that such user data will never be stored or accessible by Apple.
Integration with ChatGPT
Apple announced a partnership with OpenAI to integrate ChatGPT directly into its operating systems – which will be an optional feature that users can opt in or out of. While Apple Intelligence features will be free, access to advanced ChatGPT functionalities will require a paid subscription beyond the free quota.