Apple may be known for its calm and collected approach to new technology, but even the most calculating tech giants know a real trend when they see one. In this particular case, that trend is generative AI, aka the tech powering sophisticated chatbots like OpenAI’s ChatGPT.
According to a report from Bloomberg’s Mark Gurman, Apple is quietly developing its own rebuttal to ChatGPT called Ajax, which is being referred to internally by some engineers as — you guessed it — “AppleGPT.”
Is it surprising that Apple might be slowly wading into the seemingly irresistible waters of generative AI? Not at all. I’m sure shareholders practically demand it at this point. But that doesn’t make the prospect of an Apple-owned AI chatbot any less impactful. This could be a big moment for Apple, but an even bigger one for the ascension of generative AI.
AppleGPT
So what is “AppleGPT” exactly? Well, the short answer is (according to Mark Gurman’s intel, at least), we don’t know. It’s a large language model (LLM), to be sure — the same thing that powers ChatGPT and Google’s Bard — and Apple reportedly has more than one team working on Ajax.
But how Apple plans to use an AI chatbot, when it plans to use an AI chatbot, and just what it will be able to do are all open questions. There’s reportedly no launch date in place yet, but that doesn’t make even the mere rumor of its development any less pivotal.
For one, there’s the matter of scale. Apple is still the biggest tech company out there (actually, with its $3 trillion valuation, it’s the biggest company in history. Period). There are an awful lot of iPhones in the world, and if an AI chatbot finds its way into the iOS ecosystem, it would arguably put LLMs in the hands of more people than ever.
I know what you’re thinking, and yes Google search is also kind of a big deal. But for now, people are still mostly reliant on regular search sans chatbots. And for good reason. Chatbots (ChatGPT included) while impressive, still have a long way to go until they can be fully trusted to deliver accurate information on a consistent basis. And for that reason alone, using them for something as crucial as internet search can and should be taken with a heavy dose of “please, for the love of God don’t do that yet.”
Here’s a recent example of exactly what I’m talking about. Note: the answers provided by ChatGPT (which I used via Microsoft’s Edge browser) are all wrong.
Outside of the context of needing to deliver accurate and trusted information, however, Apple — unlike Microsoft and Google — has more wiggle room, and also more potential. I won’t feign to know if chatbots will actually wind up in iPhones everywhere, but if one were to grace such a device, I could see it generating texts in iMessage or maybe even improving the accuracy of Apple’s Siri voice assistant.
Then, there’s the Apple effect, for lack of a better phrase. While Google and Microsoft are titans of technology in their own right, Apple’s presence tends to lend credibility in a way that no other entrant can. Take the Vision Pro, Apple’s mind-bendingly expensive spatial computing headset.
While it’s still early days for the Vision Pro, just the mere presence of Apple in a space can (and already has) added a sphere of legitimacy that other tech giants can’t muster. Can Apple do that with generative AI or AR for that matter? It’s hard to tell, but it hasn’t exactly failed in its endeavors to popularize a product — however unproved — just yet.
The Pitfalls of AppleGPT
As adept as Apple is at navigating a successful course through the future of technology, there are unique problems with generative AI. For one: privacy.
As we’ve seen with recent scrutiny of OpenAI, the FTC is already attuning itself to the potential for privacy pitfalls of chatbots. According to a recent letter from the commission, a probe will assess whether Open AI has “engaged in unfair or deceptive privacy or data security practices or engaged in unfair or deceptive practices relating to risks of harm to consumers, including reputational harm.”
That’s not to say chatbots can’t be made with privacy in mind, but for a company like Apple, which still champions user privacy as a pillar of its products, rolling out generative AI that relies on massive troves of data will be a bit of a balancing act.
Even finding a way to incorporate generative AI that rises above a gimmick can be a challenge. Utility in internet search is obvious (if dangerous), but what will generative AI do on an iPad? What about on a MacBook? What utility will it provide iOS’ UX? The applications are expansive, to be sure, but not every single one is worth pursuing.
This is all to say that Apple’s potential involvement with generative AI has a long way to go before it becomes something tangible. Tim Cook himself said earlier this year that the company is being “deliberate and thoughtful” about how it might apply a chatbot to its hardware and software.
And for Tim Cook, that’s fine — Apple doesn’t need a chatbot. But chatbots on the other hand? Well, I don’t know if they can afford to miss the opportunity.
