Apple and AI
AI is anticipated to be a major talking point for WWDC 2024, with Apple expected to announce significant AI integrations across macOS, iOS/iPadOS and several stock apps.
AI has dominated the tech press since the breakout hit of ChatGPT in late 2022. Tech titans Microsoft and Google are in an arms race for business and consumer attention with products such as Gemini and Copilot. Smaller upstarts such as Rabbit and Humane with the R1 and the Humane Pin are productising LLMs. You could be forgiven for thinking that Apple has let the AI craze pass them by.
What is Hybrid AI, and why does it matter?
The vast majority of AI is currently carried out in the cloud, but there are huge potential benefits to a 'hybrid AI' approach. Hybrid AI can broadly be understood as running as much of the AI model and processing as much of your data locally as possible and resorting to using the cloud only when necessary. Many of the benefits of this approach line up well with Apple’s stance on performance, data security and privacy.
While Apple may not have rushed to launch a generative AI product, they have not been standing still on AI or "machine learning" as they've previously referred to it. Apple's silicon team have been integrating AI with their SoCs (System on a Chip) featuring dedicated 'AI controllers’ called Neural Engines, since 2017. Appearing first with the A11 Bionic in the iPhone X, and on Macs from the 2020 M1-powered MacBook. Far earlier than any other manufacturer. Since then, the number of Neural Engines inside every new Apple SoC has increased with each release. The just-announced M4 is the most powerful hybrid AI processor publicly available.
As more businesses and consumers discover AI solutions can offer time-saving workflows and new ways of working, the benefits of hybrid and on-device AI will become a key consideration in device purchasing decisions. These decisions will be made with performance, data security and privacy that only on-device and hybrid AI models can offer.
What is an 'AI-capable PC'?
In it's 2023 report analysing the current and future effects of AI-capable PCs on the market, Canalys define this new hardware category as:
"...an AI-capable PC must be a desktop or laptop possessing a dedicated chipset or block to run on-device AI workloads"
This definition could change as these devices mature and customer requirements change over time but Canalys forecasts that 19% of PCs shipped in 2024 will be AI-capable under its current definition. They speculate commercial adoption is expected to grow rapidly from 2024 onwards and that up to 60% of PCs shipped in 2027 will be "AI-capable".
So, where does Apple fit into this?
AI-capable PCs are defined by having AI-dedicated chipsets and Apple have been shipping Macs that match that definition since 2020.
The world's best-selling 13- and 15-inch laptop, the MacBook Air with M-series chips all feature advanced Neural Engines. The newly announced M4 chip, debuting in the 2024 iPad Pro has the world's most advanced on-device AI processing. Tim Cook has repeatedly asserted that machine learning, in other words, AI, has a crucial role in Apple's product roadmap.
When you combine this with Apple devices' class-leading battery life, and their relentless focus on data privacy and security, they are well-placed to lead in hybrid and on-device AI.
This focus doesn't exclude Mac and iOS users from participating in the current cloud-based generative AI scene. Mac, iPad and iPhone are also fully compatible with Microsoft Copilot and Google Gemini AI platforms as the majority of the processing is carried out on highly specialised hardware in data centres around the world. Your Mac, iPad or iPhone simply runs an app that connects you to those data centres, or 'the cloud' as it's generally known. Apps that already leverage machine learning have the best hardware to do so on an Apple device, as all current devices have dedicated hardware (Neural Engines) on the device.
As Canalys stated in their 2023 report - "Apple has AI at the core of its Macs, being the only mass-market AI-capable PC since 2020".
Is AI available today in macOS or iOS/iPadOS?
It certainly is. Chances are, you’ve been using it for a while and not noticed!
Some examples are:
- AutoCorrect now uses AI language models for more precise text predictions with MacOS Sonoma and iOS 17
- Live Text in iOS and macOS recognises text in both images and videos, making them available for user interaction (such as copy/paste and translation)
- Visual Look Up uses machine learning to allow users to learn more about the content of a photo, from identifying plants to popular art and world landmarks
- Centre Stage, Portrait Mode and Voice Isolation (for filtering out background noise on mics) all leverage AI
While Apple has used all of the above for "wow" moments in product launch and demonstration videos, they've let the feature "do the talking" and not labelled it as "AI".
So, what's next?
In the short term, Apple’s expected focus on AI at this year’s WWDC will very likely spur AI feature adoption from third-party developers across the Apple ecosystem. On-device AI processing from Apple will likely be the most secure and private way of using AI regularly. There is some speculation that Apple may create dedicated data centres, powered by M-series processors for cloud-based AI, which could be extremely interesting too.
Standard PCs and Chromebooks cannot offer local or even hybrid AI since they lack the hardware required for on-board processing. This is a huge point of differentiation compared to current Apple products. Until PC, tablet and smartphone manufacturers can integrate their version of the Neural Engine, Apple has a clear technical differentiator. Based on the recent launch of the MacBook Air M3 and iPad Pro M4, expect Apple to leverage this unique hardware advantage.
Customers using apps such as Logic Pro and Final Cut Pro will also benefit from new AI features that Apple has begun to introduce with more likely to come in future versions too. 'Pro' creative users, will see similar new features with third-party software such as Adobe CC and Pixelmator Pro which have already begun to leverage AI in their latest releases.
Businesses and education institutions, particularly those using older Windows 10 PCs (as it reaches end-of-life in October 2025), could be swayed towards Mac at upgrade. The affordability of Macs compared to the upcoming hybrid AI PCs (it is expected that hybrid AI PCs will cost 10-15% more than current models) and Apple's proven track record for data security and privacy will be compelling for companies wary of confidential data being uploaded to data centres outside of their control.
No Comments yet. be the first to comment.