Apple Intelligence is cause for both celebration and concern

Apple is not a company that takes big risks or truly innovates all that often, certainly not as often as it would have everyone believe. Allow yours truly and every other tech reporter that knows this, then, a moment to enjoy the spectacle of this particular company taking a huge risk now in the field of artificial intelligence so as to satisfy its shareholders, excite consumers and, frankly, stop looking like a distant sixth in the race every other Silicon Valley giant has already joined.

Whether Apple should have taken that risk earlier, whether it’s now taking it under favorable circumstances or whether it’s already too late remains to be seen. For the time being, let us celebrate this choice for what it is: a bold step forward when the company needed it the most, as well as proof of the fact that even Apple can give in to peer pressure if its perceived value (the only value Wall Street cares about) is threatened.

The huge risk the company is taking with AI – by bringing these generative models and machine learning algorithms to hundreds of millions of devices with the release of iOS 18, iPadOS 18 and macOS 15 – is of two kinds: the short-term, obvious one and the long-term, potentially more important one. Here’s why.

The short-term risk: ChatGPT

The obvious, short-term risk Apple is taking relates to the integration of Siri with ChatGPT, as it was widely rumored and indeed demonstrated during the company’s event. OpenAI’s famous chatbot may have come a long way since December 2022 and the drama surrounding it last year may have died down, but the technology powering it is still imperfect, still unpredictable, still capable of producing misleading or untrustworthy results, being – by nature – in a constant beta state.

AI enthusiasts will claim that ChatGPT 4o is by no means as hilariously or as dangerously wrong as often as it was in the early days (which is true). They will also claim that most people using OpenAI’s chatbot often enough are aware of all this, so they understand that double-checking and verifying every prompt’s results before using them is basically part of the process (which is also true). This, though, simply can’t be taken for granted for the hundreds of millions of mainstream consumers ChatGPT will be suddenly accessible to when the latest versions of Apple’s operating systems for the iPhone, iPad and the Mac are released.

The problem with this – as Microsoft can attest to – is that, if any of ChatGPT’s answers prove to be awfully wrong about something important, leading to serious consequences for people just blindly trusting its answers, it’s not OpenAI that will be taking the blame. It’s Apple. It will be Apple’s fault in the eyes of the public, even if the company has no control over the answers OpenAI’s servers will be offering to iPhone, iPad or Mac owners.

Apple knows this. It’s the reason why it made quite clear that Siri’s ChatGPT integration and Apple Intelligence – its own machine learning-powered functionality that’s added system-wide to its latest operating systems – are two different things, meant to serve different needs in different situations. The company has even added a “Check important info for mistakes” note under every result produced by a ChatGPT prompt, presumably linking to a page explaining that the information provided in each answer may not be accurate and that Apple should not be held responsible for that.

Yours truly does not know enough about legal matters to be sure whether that small note gets Apple off the hook in case of anything undesirable happening because of ChatGPT’s results appearing on the screen of its devices. But he does know enough about Apple’s obsession with its public image (as does everybody else). Hence the obvious risk the company is taking with ChatGPT at the fingertips of so, so many people come the fall.

The long-term risk: consumer indifference

The other kind of risk, the long-term one, has nothing to do with ChatGPT but it could prove to be equally important. Maybe even more so. Based on what Apple showed during its presentation – and even what many of the company’s own executives seemed to be concerned about a while back – there’s always the possibility that consumers may just not be all that impressed by what Apple Intelligence functionality brings to the table.

The likelihood of this happening is not low. The demos shown during Apple’s presentation were rather specific and mostly controlled, meaning that the experience of using these new AI-based functions day-in, day-out may differ wildly for consumers. Some of those looked interesting and genuinely useful, some others decidedly less so. We will not know for sure until Apple’s new operating systems are actually released, but unless the company has seriously underpromised in order to overdeliver here, anything less than what was shown in its showcase would be considered a disappointment, no?

Apple knows this too. That is why it will not be releasing every new Apple Intelligence function at once: a number of them will not be available at launch and a few among those may well slip into 2025. The problem with this approach is that it might seem too much like Apple Intelligence is going through a prolonged beta period (never a good look) while also bringing back memories of the failed Siri, which was also not ready for prime time when released (the fact that it did not evolve into anything impressive since did not help either).

The worst part in all of this is that, in the tech world, products and services really do only get one chance to make a first impression – and Apple can’t really afford to make a bad such impression now, as it’s already considered to be late in the AI game compared to, say, Google or Microsoft.

Yes, the company has a number of good ideas about how machine learning and generative models can fit into the everyday use of its devices, but it has to nail the implementation part this time around. Its loyal customer base may show more patience than usual – affording the company a bit of time – but all eyes are now on Apple because, simply put, this has been a long time coming and expectations are almost unfairly high. It does seem like a double-edged sword right now, all that big talk, no?