During the first hour of Apple’s Worldwide Developers Conference (WWDC), “intelligence” was mentioned only three times. However, in the next 40 minutes, it was hard to keep track when Apple unveiled its rebranding from “Artificial Intelligence” to “Apple Intelligence . ”
A Quick Summary
Apple has announced a minor partnership with OpenAI , relying primarily on its own models for many of the new features. Tim Cook, CEO of Apple, highlighted that Apple Intelligence will be “powerful, intuitive, integrated, personal and private.” The features will be released this fall, and it remains to be seen how effective they will be in terms of power and intuition.
AI integration
Apple Intelligence will be integrated into three main areas: writing tools, image generation, and a Siri that performs actions. These features can be described as interesting, average and surprising, respectively. All activity on the device will be stored in a semantic index that personalizes the AI. Apple Intelligence will select between on-device and cloud models, ensuring cloud computing is private.
Apple Models and Partnership with OpenAI
Apple uses three levels of Long Language Models (LLMs) for these features:
- Apple On-Device: A 3 billion parameter model, similar in performance to the Gemini 1.0 Pro.
- Apple Server: A larger model hosted on Apple’s private servers, with performance possibly similar to Claude Sonnet.
- Third-party models: Starting with GPT-4 from ChatGPT, although models similar to those of Google Gemini are expected.
Apple Intelligence decides which model to use based on the user’s request. Before interacting with ChatGPT, it will always ask for permission.
Siri integration
With the OpenAI partnership, Siri can interact with ChatGPT without needing to create a ChatGPT account, although an existing account can be connected if desired. This integration is limited initially, but could be expanded to improve Siri’s capabilities in the future. Full integration with OpenAI is expected later this year.
Siri 2.0 Features
Siri has received a significant update thanks to Apple Intelligence. Now, you can take actions across all apps, which is a significant improvement. Siri will also become smarter about the user’s personal context, thanks to a semantic index that stores all activity on devices.
New Siri features include:
- Screen Awareness: Siri will be able to understand and interact with content on the user’s screen.
- Product knowledge: Siri will become an expert on Apple devices, guiding users even without knowing the exact name of the feature.
Writing Assistance
Apple Intelligence will introduce a small typing assist window in all applications. This tool can review and rewrite text with built-in presets such as friendly, professional, and concise, or convert existing content into summaries, key points, tables, or lists. Plus, it suggests automatic replies and summaries for emails, calls, and voice memos.
Image Generation
Imaging is coming to Apple devices, limited to styles like animation, illustration, and sketch. On-device models create images based on the elements and styles selected by the user, integrating into messages, notes and personalized emojis (GenMoji).
AI in Underlying Functions
Apple has implemented AI in underlying features like the new Calculator on the iPad, which performs interactive calculations based on user input. Also, language models enhance focus and productivity features, and image models allow you to create automatic memory movies, search through photos and videos, or remove unwanted objects from images.
Is Apple Intelligence a Big Advance?
Apple Intelligence promises to make life easier for users by making AI accessible to the masses. While the on-device models may not impress power users, these new features will significantly improve the average user experience.
The new Apple Intelligence features will be available with upcoming versions of the operating system for iPhone Pro and iPhone Pro Max, Macs with M1 chips and later, and iPads.
Conclusion
Apple is taking significant steps to integrate artificial intelligence into its devices, seeking to offer a more personalized and private experience. The real test will be to see how these new features perform in the real world when they launch this fall.