Another year, another Apple WWDC around the corner—for Apple, what matters most isn’t just Tim Cook’s farewell speech, but answering the world’s expectations for AI.
Apple must confront its most embarrassing issue in the past three years—why does the world’s most expensive smartphone come with the dumbest AI assistant?
On May 28 local time, ten days before the launch event, foreign media broke the answer.
It is reported that this overhaul of Siri is the largest since its debut with the iPhone 4S in 2011. The new interface features a dark color scheme, has been rebuilt around a chatbot interaction model, and is deeply integrated with Dynamic Island.
More importantly, Apple will allow users to directly “insert” Google Gemini and Anthropic’s Claude into the Siri experience—Siri is becoming a distribution platform for AI models.
Everyone is curious: what will Siri look like with AI integrated?
01. Completely redesigned interface
According to Bloomberg, the new version of Siri has several core changes that, when viewed together, reveal Apple's comprehensive strategy.
Possible integration interaction between Siri and Dynamic Island | Image source: Instagram
The first is a complete overhaul of the interface: a chatbot-style interaction, dark color scheme, and Dynamic Island integration—Siri has transformed from a "popup" into a standalone app experience entry point. This is not just a visual upgrade; more importantly, it signals that Apple wants users to treat Siri as a tool they actively use, rather than a voice command accessed occasionally.
The second is conversational persistence. For years, one of the biggest frustrations with Siri has been its inability to remember context—each interaction starts from scratch, with no memory or continuity. The new version of Siri is reportedly fixing this issue—it may sound minor, but it’s a foundational requirement for a true assistant experience.
Third, and most notably, is the "Extensions framework"—which allows third-party AI models to integrate with Siri.
The deeper intention behind this design is that Apple is no longer treating "building the best AI model" as the only path forward; instead, it is repositioning iOS as "the platform on which the best AI models compete." Just as the App Store doesn’t require Apple to develop all apps itself, the new Siri ecosystem doesn’t require Apple to outperform everyone in model capabilities—it simply needs to bring together diverse models and retain users through system-level integration.
In short, Apple is fighting a model battle with a channel mindset.
02. Siri's Three-Year Debt
To understand the significance of this update, you first need to recognize how passive Apple has been over the past few years.
In 2023, ChatGPT emerged and redefined what “conversational AI” means. In 2024, Google integrated Gemini into Android, and Samsung made AI features a key selling point. The entire industry is racing forward—but what is Siri doing? It still misunderstands user commands, turning “Set an alarm for 8 a.m. tomorrow” into simply opening the alarm app.
Apple, of course, did not sit idle. At WWDC 2024, Apple Intelligence made a grand entrance, promising a host of deeply integrated AI features. But in reality, many of these features have been delayed, restricted to specific regions, or fall far short of the demos presented on stage. A long-time Apple tech analyst bluntly stated: “This doesn’t feel like a polished comeback—it’s more like Apple finally joined the AI race, only to realize it’s still in mid-development.”
After three years of decline, Apple urgently needs a true turnaround.
Two days ago, Apple quietly launched the subdomain genai.apple.com. This subtle move has sparked considerable buzz in the tech community, with many interpreting it as a signal that Apple is making its final public preparations for the "AI transformation" at this year's WWDC.
03. The Essential Questions to Answer
But here lies a paradox that many media outlets have already discussed.
One of Apple’s most important long-term moats has been privacy. “Your data is processed only on your device” is Apple’s core promise to users—and the reason behind the Private Cloud Compute architecture.
Now, to make Siri stronger, Apple is integrating Google's infrastructure to handle certain AI queries.
This is not a technical issue; it’s a trust issue.
Possible Siri Q&A interface | Image source: Instagram
When Apple itself breaks the红线 of "using only its own computing infrastructure," its commitment to user privacy is no longer absolute. Users can, of course, choose not to use Google Gemini integration, but "can choose not to" is entirely different from "defaultly untouched." How Apple explains this shift to users during the keynote will be one of the most important details to watch on June 8.
Moreover, there is a more fundamental issue. A user on Reddit asked a simple but pointed question—If the experience with Claude in Siri is the same as using Claude directly, why should I use the wrapped version?
Apple must provide a compelling answer, and currently, there is only one candidate: system-level integration—an AI that can access contacts, calendar, photos, and health data offers a completely different experience from an AI running in isolation.
This is Apple's final and most important card.
There has been much criticism of Apple’s pace in AI, but a counterintuitive logic is also circulating—that Apple may be slow because it’s waiting for others to stumble first.
Over the past two years, OpenAI, Google, and Meta have invested hundreds of billions of dollars in data centers, chips, and model training, sparking concerns about an AI bubble. In contrast, Apple’s strategy is to avoid rushing to build the “most powerful model”; instead, it waits for the赛道 to stabilize, then leverages its core strength in system integration to catch up and surpass the competition.
To some extent, this iOS 27 layout is fulfilling this logic: instead of competing on model capabilities itself, it integrates Gemini and Claude, then builds a differentiated user experience moat using features like Dynamic Island, personal data permissions, and on-device processing—capabilities Android cannot replicate.
This is not a desperate rush by latecomers; it is a calculated bet.
The bet is that the endgame of AI isn't about whose model is the strongest, but whose system uses the model most effectively.
On June 8, Apple will provide its full answer. That will be the true test of this gamble: whether Siri can truly win over users already accustomed to ChatGPT and Gemini.
Fifteen years later, Siri owes its users an explanation.
*Header image source: Instagram
This article is from the WeChat public account "GeekPark" (ID: geekpark), author: Hualin Wuwang.