GAFAM NewsEssay

Is Siri AI late, or deliberately last? Reading Apple's WWDC26 bet

Apple unveiled Siri AI at WWDC26. Is it a catch-up move or Apple's classic late entry? What it borrows from Google and what it must prove this fall.

Read the Japanese original on note「Siri AIは遅れか、後出しか」

Good evening. The Owl here.

This week in GAFAM news, Apple’s WWDC was the main story.

At WWDC26, Apple announced the next generation of Apple Intelligence and a new “Siri AI.” Apple describes Siri AI as a new assistant that understands what’s on your screen, draws on personal context from your email, messages, photos and more, and takes actions across apps. [1]

Honestly, if you only read the headlines, the feeling is “Well, it’s finally here.” ChatGPT, Gemini and Copilot set the mood of the market first. Google turned search itself into AI, and Microsoft started putting AI agents at the front door of everyday work. Against that backdrop, Apple seems to have rebuilt Siri very cautiously, or simply late.

But Apple has never been a company that ships first.

It’s a company that’s good at going last. To put it more precisely, it’s good at watching others clear a rough path into a market, waiting until the technology matures, and then pulling together hardware, the OS, apps, privacy, billing and its developer ecosystem to fit the whole thing into everyday life.

So is this Siri AI a belated attempt to catch up? Or is it a classic Apple late entry, made only once Apple was good and ready?

Let me try to read it, borrowing some of the public reaction along the way.

What Apple brought out wasn’t a chatbot. It was AI inside the OS

The first thing to note is that Apple isn’t presenting Siri AI as just a chatbot.

According to Apple, Siri AI is deeply integrated into iPhone, iPad, Mac, Apple Watch and Apple Vision Pro. It is designed to understand what’s on the user’s screen, find the information it needs in personal email, messages, photos and so on, and even carry out actions inside apps. [1]

That’s a little different from an AI like ChatGPT, where you “go somewhere else to ask for help.”

What Apple is aiming for is a state where, before the user even opens an AI, the AI is already waiting within the context of the OS. Calendar, Photos, Mail, Safari, Messages, Shortcuts. Siri AI seeps into the places people already use every day.

The Verge’s hands-on article watched the new Siri add events by looking at email and the calendar, and use information on the screen to move on to the next action. Its verdict was that these are fairly basic AI assistant features, but the fact that they actually work as Siri is a big deal in itself. [2]

This is where it gets interesting.

Across the AI industry, the conversation tends to center on how smart a model is, benchmarks, reasoning ability, video generation and agent performance. For Apple, though, showing off a cutting-edge model on its own matters less than “working naturally, and rarely failing, on the device you use every day.”

Apple’s AI seems to be competing not through flashy research announcements, but on how seamlessly it fits into the paths people already take through their daily lives.

But the public reaction was pretty cool-headed

That said, the reaction was not unreservedly positive.

Another piece from The Verge sees Apple’s announcements this time as bold AI promises that were, for the most part, about catching up. Apple stepped into AI agents in earnest, but its direction (multimodal chat, screen understanding, personal context, actions in apps) covers territory that Google, OpenAI and Microsoft had already been moving into ahead of it. [3]

In other words, there are two ways to look at this Siri AI.

One is that “Apple has finally made it back to the starting line of the AI era.”

The other is that “Apple didn’t rush onto everyone else’s field. It’s trying to pull AI onto its own field.”

I think both are right.

For a long time, Siri was weak compared with what people expected of it. It was a pioneer among voice assistants, yet at some point it became “something you can use for timers and the weather, and that’s about it.” Then the generative AI wave arrived, and Apple clearly fell behind.

But falling behind doesn’t necessarily mean it’s over.

What Apple has is devices that people around the world touch every day, and permissions at the OS level. It doesn’t need anyone to install a new AI app. It doesn’t need anyone to log in again. Apple is already sitting right next to users’ photos, messages, email, calendars, location data and app activity.

So Apple’s path to winning may not be “the smartest AI,” but “the AI that can work its way furthest into everyday life.”

Borrowing Google’s strength without turning into Google

Another thing that drew attention this time was Apple’s relationship with Google.

According to 9to5Mac, Craig Federighi talked about Apple’s collaboration with Google on Siri AI in his briefings after WWDC. But Federighi explained that Apple isn’t simply using the Gemini app, Google’s consumer models or Google Search as they are. He stressed that Siri AI runs on Apple Intelligence system software, on-device models and Private Cloud Compute. [4]

There’s a very Apple-like line being drawn here.

When it comes to foundational AI technology, it isn’t easy for Apple, on its own, to catch up head-on with Google, OpenAI, Anthropic and Microsoft. So it borrows outside strength. But it keeps control of the user experience and the privacy design on Apple’s side.

That’s close to something Apple has done many times before.

In many areas, including components, connectivity, cloud, search, maps and payments, Apple has relied on outside players. But the experience the user sees is made into Apple’s own. With Siri AI too, Apple is borrowing Google’s technology while trying to present it not as “We put Google’s AI on the iPhone,” but as “We rebuilt it as an Apple AI experience.”

This is both a strength and a risk of Apple going last.

The strength is that users don’t have to think about who supplies the AI or what the model is called. The risk is that the more Apple relies on another company’s models behind the scenes, the more people will ask, “So where is Apple’s own AI capability?”

Apple’s chosen differentiator: a “curt” AI

What I found interesting was the reaction to Siri AI’s personality.

The Verge actually praised Siri AI for giving short answers and for not trying to drag the conversation out. Whereas many AI chatbots are friendly, wordy and eager to keep you talking, it wrote, Siri AI is closer to a tool that gives you only what you need. [5]

It’s a low-key choice, but a very Apple one.

Many AI companies have tried to present AI as a “partner” or a “friend.” The conversation keeps going; the AI empathizes, encourages, and sometimes creates a feeling close to a human relationship. That can be useful in some situations, but it also brings problems of dependence, misunderstanding and excessive anthropomorphism.

Apple seems to be keeping a bit of distance from all that.

Siri AI is not your friend. It’s your tool.

If Apple really sticks to this direction, it’s more than a personality setting. The deeper a company puts AI into people’s lives, the heavier its responsibility to design the distance between people and AI. When an AI that reads your personal email, photos and calendar also behaves like an emotional companion, the line between convenient and creepy gets blurry.

Perhaps Apple is deliberately making it a little cold and practical for exactly that reason.

The biggest challenge isn’t shipping it, but winning back trust

Still, Siri AI comes with some big conditions.

According to Apple’s official announcement, testing of Siri AI has begun for developers, but for general users it is scheduled to arrive as a beta later this year on supported devices set to English. It also won’t be available at first on iPhone, iPad and Apple Watch in the EU, and it won’t be available in China either, because of regulatory requirements. [1]

In other words, being announced doesn’t mean everyone can use it right away. This matters a lot for Apple. That’s because, around Apple Intelligence and Siri, there’s a lingering memory of past promises that were slow to materialize. The Verge’s hands-on article also sees two goals Apple can’t afford to miss with this Siri: getting it to actually work, and getting it into users’ hands. [2]

What’s being tested here isn’t how well Apple puts on a keynote. It’s trust.

Apple isn’t a company that shows off new technology to the fullest the moment it exists. If anything, it’s a company that waits until something holds together as an experience and then rolls it out in a usable form. That’s why people expect that “it’s slow, but when it ships, it really works.”

This Siri AI is a test of whether Apple can win that expectation back. If, in the beta from this fall onward, actions across Mail, Calendar and Photos run reliably, users are convinced by the promises of on-device processing and Private Cloud Compute, and it genuinely cuts down the small hassles of daily life, then Apple’s late entry will have been well worth it.

If, on the other hand, the impression grows that it’s amazing in demos but shaky in real life, that regions and languages arrive slowly, and that the device requirements are strict, then it will look less like “good and ready” and more like “belated.”

This week’s take: Apple didn’t flee the performance race for an experience race. It moved the race

Watching this WWDC, this is what struck me most.

Apple hasn’t run away from the AI competition.

But it is trying to shift where the competition takes place.

While Google and OpenAI put “how smart a model we can build” front and center, and Microsoft pushes “how much work you can hand over to AI,” Apple is going after “how naturally AI can melt into the device and into daily life.” That is Apple going last, in classic Apple style.

This time, though, Apple doesn’t have much room to take its time. The bar for AI assistants has already been raised. Through ChatGPT and Gemini, users have already learned how much “you can get done just by asking an AI.” Apple can use privacy and an integrated experience as weapons, but those alone won’t earn it forgiveness for weak performance or late delivery.

Apple’s Siri AI looks less like a flashy, come-from-behind knockout and more like a first step toward rebuilding trust. And for Apple, the real contest isn’t the applause at WWDC. It’s whether, when users pick up their iPhones this fall and beyond, they think, “Oh, I might actually use this every day.”

🦉 From where I sit in my tree, Siri AI is both “Apple arriving late” and the beginning of “the late entry Apple is so good at.” The next question is whether it can become an AI that doesn’t show off how smart it is, but quietly takes the hassle out of daily life.

References

[1] Apple Newsroom: Apple unveils next generation of Apple Intelligence, Siri AI, and more

[2] The Verge: I tried Siri AI, and so far it actually works

[3] The Verge: Apple’s AI promises are finally, almost, sort of here

[4] 9to5Mac: Craig Federighi details Apple’s collaboration with Google for Siri AI in iOS 27

[5] The Verge: Apple’s new Siri AI knows when to shut up