The genius of Apple’s AI announcement

This week Apple finally unveiled its response to the generative AI explosion kicked off by ChatGPT in November 2022.

After its recent misstep with the ‘Crush!’ iPad ad creative, it’s a reminder that when Apple brings its product, comms and marketing A-game, nobody does it better.

Here are 8 ways in which it nailed it on Monday:

1.) Appropriating the initialism

Having hitherto shunned the AI label, Apple has now boldly appropriated it, replacing the word ‘Artificial’ (which gives many people the heebie jeebies) with ‘Apple’ (one of the world’s most valuable brands). There was just two mentions of ‘artificial intelligence’ (no caps) in the keynote: one making it clear Apple have “been using artificial intelligence for years” and one positioning Apple’s approach as “beyond artificial intelligence”.

2.) Positioning for the mainstream

A classic market positioning move but executed with aplomb, the strapline for Apple Intelligence - ‘AI for the rest of us’ - simultaneously paints current AI services as for a niche and exclusionary group and Apple’s as for the mainstream (never mind that you’ll need an £1,000+ A17 Pro iPhone or a £500+ Apple silicon iPad/Mac to access it).

3.) Differentiating on personal utility

Whilst Monday’s demos lacked any jaw-on-the-floor moments, there was an accumulation of a large number of useful features. At a time when Microsoft seems to be increasingly prioritising its enterprise customers (e.g. jettisoning GPT Builder for non-enterprise users), Apple is anchoring around the value generative AI can bring to the individual (in the hope it will keep them in/bring them back to the Apple ecosystem, with its hardware replacement cycle, service subscriptions, App Store commission and, increasingly, ad-revenue).

4.) Leveraging control of the OS

Apple’s system-wide AI writing tools

Many of the AI-powered features Apple demoed we’ve seen in standalone AI products in the last 18 months. Weaving those features throughout Apple’s operating systems and product set (and overlaying them on top of third party apps) is going to expose them to many more people than would find their way to even the most compelling standalone AI product. It’s also going to put a lot of start-ups out of business, as standalone AI products are going to have to be really good to compete with free and more easily accessible. It has implications for Microsoft too, which is currently charging for these sorts of features within its 365 productivity suite.

5.) Differentiating on trust/privacy

Craig Federighi, Apple’s Senior Vice President Software Engineering, making other AI companies’ handling of your data sound proper dodgy.

Using every tap/swipe/keystroke/cursor movement to create more personally relevant on-device AI responses is a high-stakes game and Microsoft showed how not to do it with the rushed unveiling and subsequent rearchitecting of Recall. However, Apple has been investing in ‘trust’ for many years and I would be surprised if they messed up the implementation of this. People who’s technical assessment I respect seem to think Private Cloud Compute looks sound.

6.) Sidestepping the landmines

Generative AI has a number of major landmines (deepfakes, bias, hallucinations, training data consent), which Google and OpenAI seem to keep stepping on. By limiting image generation to non-photographic styles, Apple has deftly sidestepped the deepfake landmine and reduced the risk of its output closely mirroring an artist’s style or IP. By using smaller - and one suspects more judiciously selected - training sets for its own models and handing off to ChatGPT for tasks requiring a frontier model, it’s reduced the likelihood of consternation and litigation from creators.

7.) Reviving Siri

Microsoft and Google both opted for killing their voice assistants (Cortana and Google Assistant) and launching new ones under their AI umbrella brands (Copilot and Gemini). Apple has opted to breathe new life into Siri. Yes, the bar on improving Siri was very low, but it does look like she has comfortably leap-frogged Alexa and is now snapping at the heels of Google’s Gemini and OpenAI’s GPT-4o. Whilst it doesn’t look like Siri’s quite caught up with the voice capabilities of GTP-4o, that may not matter when your parent owns the operating system that a billion people have in their pocket.

8.) Making my predictions look prescient ;)

From my February 2024 predictions:

I anticipate Apple will use the launch of iOS 18 and the iPhone 16 to introduce a raft of new AI-enabled features, including an-LLM powered Siri.

It will focus on on-device processing, foregrounding privacy and security whilst continue to shun the AI label.

Apple’s control of both hardware and operating system on Mac computers will enable it to integrate AI in a way that is challenging for Microsoft, who are currently trying to persuade OEMs to add a Copilot key to PC keyboards.

And from my June & July 2023 predictions:

Personal

Apple’s decision to market itself as the privacy-first company you can trust with your data puts it in pole position to be the custodian of your future personal AI (sorry, personal on-device machine learning).

Integrated

Whilst some of these tools will be given their own shop window (e.g. Firefly, Image Creator), the main focus will be on integrating these new capabilities into existing products and services (e.g. Photoshop, Microsoft 365) to increase retention and/or generate a new revenue stream.

On-device

Integration at a device-level may prove even more significant and is no doubt where Apple will be placing the majority of their generative AI chips (both metaphorical and physical).

As well as avoiding cloud processing costs and alleviating privacy concerns, on-device processing can enable offline generation and a tighter integration with the operating system.

They even borrowed some of my headings for their benefits slide…