Apple’s annual conference, WWDC 2024, unveiled the zenith of its recent AI pursuits: Apple Intelligence. But what exactly is Apple Intelligence? Think of it as Apple's rendition of Windows Copilot, seamlessly woven into the fabric of all Apple devices—from iPhones and iPads to Macs. This innovation also brings ChatGPT into the mix, thanks to Apple’s alliance with OpenAI, which we’ve previously delved into.
To cut to the chase, here’s what Apple Intelligence brings to the table:
AI-Augmented Writing Tools: Powered by ChatGPT, this feature promises to elevate the linguistic prowess of Apple products. Your device will now be able to summarize lengthy lecture notes or conversations, rewrite text to match a desired tone, proofread, and even generate replies—all with a simple tap.
Enhanced Contextualization for Emails and Notifications: Whether it’s due to ChatGPT or Apple’s own language model, the outcome is clear: your Apple device will intuitively prioritize notifications and emails based on your interactions across various apps.
Integrated Image Generation and AI Recognition: Users can now generate and edit images and videos through text or sketches. Additionally, searching your photo and video library via text descriptions becomes a breeze.
An Upgraded Siri: With Apple Intelligence, Siri’s language capabilities receive a significant boost, incorporating contextual understanding of your data and interactions. Siri will now seamlessly interact across different apps.
In-App User Interface for Apple Intelligence Writing Tool
While these features may not be groundbreaking for seasoned users of generative AI tools and Copilot, they promise to revolutionize our interactions with smartphones—extending beyond just Apple devices. At present, only Samsung has introduced similar AI features in its Galaxy lineup, utilizing Google’s Gemini Nano for on-device tasks and Gemini Pro for more intensive, cloud-dependent tasks.
The Difference Between Samsung and Apple’s On-device AI
Samsung’s Galaxy AI was the pioneer of smartphone AI, offering AI-augmented writing tools akin to Apple Intelligence, alongside features like Live Translation for phone calls and image editing. Yet, Galaxy AI hasn’t basked in the limelight as one might expect. Why, you ask? Let’s delve into the matter:
Shortcomings of Gemini LLMs: Google’s track record with generative AI has been less than stellar (remember Bard?). Additionally, Gemini only made its debut in December 2023, a mere four months before its integration into Samsung smartphones.
Limited to Samsung Default Apps: Galaxy AI’s brilliance was shackled to Samsung’s default apps, not extending its magic across the entire Android ecosystem. This limitation meant that many users, who preferred other apps, missed out. Furthermore, the AI’s ability to understand and enhance user experiences was curtailed due to restricted access to contextual data in other apps.
As for Apple Intelligence, while the exact details of its seamless integration across iPhones remain shrouded in mystery, we can be confident that the prowess of ChatGPT will be at least on par with its current iteration, ChatGPT-4. The efficacy of Apple’s proprietary on-board language model is yet to be proven but given the substantial research and development efforts from Apple, it’s fair to say the prospects look promising.
Apple is Intent on Building an On-device AI that Surpasses All
Though Apple has yet to unveil a generative AI tool, its strides in developing on-device AI models are noteworthy. Recently, the tech behemoth released 20 new Core ML models and 4 datasets on the prominent open-source ML library, Hugging Face. This pivot towards open source is quite uncharacteristic for Apple, known for its tightly controlled ecosystem of devices and software. However, it signals Apple's recognition of the importance of developer adoption and contribution to its AI models. It underscores the company’s commitment to AI and its practical applications, suggesting that Apple Intelligence might become the first truly effective implementation of on-device AI on smartphones.
While Apple hasn’t been at the forefront of smartphone innovation in recent years, its implementations are renowned for their efficacy. In the same vein, although Samsung’s Galaxy AI made its debut first, Apple has the potential to deliver a more refined and effective generative AI experience across its devices.
Admittedly, Apple hasn’t been the swiftest in the generative AI race. However, if Apple Intelligence successfully integrates compelling generative AI features, it could position Apple as a leader in on-device LLMs. A successful rollout would see Apple’s own on-device LLMs working seamlessly with ChatGPT, offering users an AI experience that becomes an integral part of their daily lives.
Looking ahead, Apple faces a pivotal choice. It could either keep its on-device LLMs exclusive to its devices, providing a competitive edge or allow other companies to utilize its LLMs, much like OpenAI’s partnerships. If Apple chooses the latter, it could very well pave the way to becoming the dominant force in on-device AI models.
Comments