Apple Intelligence a Home Run for Apple AI?

0
Jun 20, 2024
  • Unlike Google, Apple avoided creating any grandiose narrative for AI at WWDC. Instead, it focused on specific, practical applications.
  • Apple AI’s focus is on intuitive features and seamless cross-app integration.
  • Through the introduction of Apple Intelligence, Apple focuses on harnessing its chip expertise and hardware-software synergy.

The world has been anxiously waiting for Apple’s AI roadmap for over 18 months since the rise of ChatGPT, and it has finally arrived. In this blog, we provide a quick rundown of this year’s WWDC and key analyses of Apple’s AI strategy.

Clear, practical and intuitive AI features

Unlike Google, Apple avoided creating any grandiose narrative for AI. Instead, it focused on specific, practical applications. The company has introduced Apple Intelligence, which it describes as “the personal intelligence system for iPhone, iPad, and Mac that combines the power of generative models with personal context to deliver intelligence that is incredibly useful and relevant”. The focus here is on intuitive features and seamless cross-app integration:

Text: Apple Intelligence has introduced a feature called Priority Notifications. When this feature is enabled, the system sorts notifications based on their importance. The most critical information, such as an email containing the user's boarding pass, will be displayed at the top.

Another key feature is the Systemwide Writing Tool, which is available across Mail, Notes, Pages and various third-party apps. Users can expand, proofread and summarize text in different applications. Besides, they can adjust the tone of their wording or receive suggestions for smart replies.

Image: Apple Intelligence introduces Genmoji and Image Wand features. Genmoji helps users create original images and emojis, but currently only supports sketch, cartoon and illustration styles. Image Wand can transform users' rough hand-drawn sketches into more refined images. It can also generate images based solely on text descriptions, just like Stable Diffusion.

New Siri: The highlight of this event was the upgrade to Siri. Driven by Apple Intelligence, Siri is now equipped with a “semantic indexing feature”, promising a somatic understanding of its users’ intentions. This allows Siri to organize a semantic index library based on the user’s photos, calendar, files, emails and more, enabling smarter information search and retrieval.

It is worth noting that Apple has designed a brand-new framework for Apple Intelligence. This framework allows developers to introduce Apple Intelligence into their applications through the App Intents API. This means that Siri will be able to perform complex operations across multiple apps in the future.

Last but not least, Apple has entered a collaboration with OpenAI, linking the recently released GPT-4o model with Siri. Users can access most of the model’s features for free without needing to create an account.

AI strategy analyses

In the arena of large-model innovation and AGI race, Apple’s core technology has been admittedly lagging. However, at this year’s WWDC, the company’s precise tactical approach shone once again.

  • First and foremost, Apple highlights the existence of its proprietary AI models, marking its entry into the generative model arena. Unlike Google and Microsoft/OpenAI, Apple doesn’t have a frontier LLM. However, it has developed a suite of small and medium generative models, including a 3-billion-parameter on-device language model and a larger server-based model accessible via Private Cloud Compute on Apple silicon servers.
  • Apple leverages its unique ecosystem, mitigating the impact of its temporary technological lag. Through the introduction of Apple Intelligence, it focuses on harnessing its chip expertise and hardware-software synergy, emphasizing the “on-device” large-model capabilities and addressing common concerns about data security and user privacy in large-model applications. This creates a distinct space for Apple’s AI despite its initial shortcomings.
  • Apple develops partnerships with leading large-model provider OpenAI, establishing a connection between Siri and GPT-4o. This provides users access to the most advanced models.
  • However, Apple subtly downplayed the partnership with OpenAI, only announcing it at the end of the event. It emphasized that this collaboration would require user permission for external access rather than being deeply integrated at the OS level, ensuring the advanced model doesn’t overshadow Apple’s own systems.
  • In fact, Apple is adopting a model agnostic approach. Its partnership with leading LLM providers will not stop at OpenAI. Google’s Gemini and other specialized, fine-tuned large models could be the next in line for integration with the iPhone.
  • Recognizing that many iPhone users may be new to generative AI, Apple championed “AI for the rest of us” at WWDC. This consumer-friendly approach aims to drive widespread AI adoption on Apple devices. By leveraging its unique hardware, software and private cloud infrastructure, Apple not only differentiates itself from OpenAI but also strengthens user trust and reinforces customer loyalty.

In the fast-paced developments of generative AI, all the leading players, exemplified by OpenAI and Google, are trying to “move fast and break things”. Apple’s WWDC has signaled that the slow and steady can also win the race, especially in its edge AI.

In our upcoming short report on Apple Intelligence, we will further dissect Apple’s AI and its impacts on industry ecosystems and players. Moreover, Apple’s patent applications reveal numerous machine learning-powered innovations across its devices. Securing these patents not only strengthens Apple's position in the AI devices market but also helps fend off competitors. Based on these insights, we will forecast Apple's next strategic moves.

Summary

Published

Jun 20, 2024