Editorial note: At GadgetBond, we typically steer clear of overtly political content. However, when technology and gadgets, even the unconventional kind, intersect with current events, we believe it warrants our attention. Read our statement

Apple wanted Siri AI to be one of the headline features of its next big software cycle. Instead, in Europe, it has just become a case study in how far regulators and platform companies are willing to push each other over artificial intelligence, privacy, and control of the smartphone.

On June 8, Apple quietly dropped a significant update in its Newsroom: Siri AI, its new Apple Intelligence-powered assistant for iPhone and iPad, will not ship in the European Union when iOS 27 and iPadOS 27 roll out later this year. The company squarely blames the EU’s Digital Markets Act (DMA), saying months of talks with regulators failed to produce a compromise that would let the feature launch while still keeping what Apple calls essential privacy and security protections intact.

If you are in the EU, the practical takeaway is blunt: your iPhone and iPad will get iOS 27 and iPadOS 27, but without the new Siri AI experience that Apple is heavily marketing elsewhere. That means no new Siri AI app to revisit conversations, no expanded Visual Intelligence that can understand on-screen content more deeply, no integrated AI writing tools, and no new Siri mode in the Camera app. Even Apple Watch owners get caught in the blast radius, because Siri AI on watchOS 27 relies on a paired iPhone with Siri AI enabled, so the experience simply will not be available there either for EU users.

Curiously, the situation is not a blanket EU ban on Apple’s AI push. macOS 27 and visionOS 27 will still ship with Siri AI inside the EU, so Mac and Vision Pro users in Europe will see at least some of the same “Apple Intelligence” story that their counterparts in the US or India will. The split underscores that this is not about some abstract regulatory hostility to AI as such, but about how deeply a virtual assistant can plug into a smartphone’s operating system and other apps without falling afoul of the DMA’s interoperability rules.

At the heart of the dispute is a clash of interpretations of what the DMA requires from so-called “gatekeepers” like Apple when they build tightly-integrated digital assistants. Apple argues that under what it calls the EU regulators’ “extreme interpretation” of the DMA, it would be forced to grant any virtual assistant “nearly unlimited access” to a user’s device, and the ability to act on that access autonomously. In Apple’s telling, that would mean that once Siri AI is live in the EU, any competing assistant could demand the same hooks: reading and sending messages, making purchases, accessing files, and executing actions across apps – effectively the same level of control and access Apple is giving its own assistant.

From a user privacy and security perspective, you can see why Apple is framing this as a red line. The company cites security research showing that AI systems can be hijacked or manipulated to exfiltrate personal data like passwords and photos, or silently modify files and account settings, as the risk profile of assistants grows alongside their power. The more an AI system can do on your behalf, the more catastrophic it becomes if that system is compromised, misconfigured, or simply tricked by malicious prompts.

To thread that needle, Apple says it proposed a middle path to Brussels: a concept it calls a “Trusted System Agent.” The idea is to put an intermediary layer between virtual assistants and the OS, giving third-party assistants access to the same high-level capabilities Siri AI enjoys, but with guardrails Apple believes are necessary to preserve privacy and security. Importantly, Apple offered to phase this in over 18 months, giving regulators time to examine the architecture while still getting Siri AI into users’ hands sooner rather than later.

According to Apple, the European Commission was not persuaded. The company says regulators rejected not just the Trusted System Agent idea, but “did not agree to any of Apple’s proposals,” leaving Apple in a position where, in its view, the only compliant option would be to open up deeper device access immediately to any assistant that asks. Rather than do that, Apple is opting to hold back Siri AI on iOS, iPadOS, and watchOS in the EU altogether, without offering even an approximate timeline for when or whether that decision might change.

Zooming out, this is not happening in a vacuum. The DMA is the EU’s flagship competition framework for large digital platforms, targeting what regulators see as bottleneck gatekeepers that can tilt markets in their favor. It mandates, among other things, interoperability and anti self-preferencing obligations for core platform services, including operating systems and voice assistants, and the European Commission has not hesitated to open investigations when it believes compliance falls short. Apple has already had to make substantial changes in the EU – from allowing alternative app marketplaces to loosening some NFC and browser defaults – and Siri AI is simply the latest front in that ongoing tug-of-war.

For developers in Europe, the ripple effects are significant. Apple notes that EU-based developers will not be able to test or integrate the new Siri AI features into their apps on iOS 27, iPadOS 27, or watchOS 27, even if those apps are distributed globally. That means European teams working on global apps will effectively be locked out of one of Apple’s marquee AI platforms unless they operate test environments or subsidiaries outside the EU, adding friction to an ecosystem Apple has historically prided itself on making as uniform as possible.

It also sets up a fragmented user experience. If you are in, say, the US or India, you might be using Siri AI to summarize notifications, understand what is on your screen, or automate actions across apps in a more conversational way on your iPhone. Your friend or colleague in Germany, meanwhile, will be running the same iOS version number but with a more constrained Siri and fewer AI hooks, while still potentially seeing Apple’s marketing for Apple Intelligence on other platforms. That is not the cross-platform continuity story Apple likes to tell – but it is increasingly the reality when global regulation collides with tightly-controlled ecosystems.

The politics here are messy. Apple’s statement is framed in unusually sharp terms for a Newsroom update, accusing regulators of refusing to “engage constructively” and emphasizing that there is “currently no timeline” for Siri AI’s arrival on iPhone and iPad in the EU. The European Commission, for its part, has every incentive not to be seen as rubber-stamping proprietary middle layers that could limit true interoperability, especially at a time when generative AI is quickly becoming another arena where a handful of US-based giants could consolidate power. The result is a stalemate where users are largely collateral damage.

There is also a philosophical difference playing out. Apple’s entire Apple Intelligence pitch is “private by design,” leaning heavily on on-device processing and its Private Cloud Compute infrastructure, which routes more complex AI workloads through Apple-controlled servers designed with end-to-end privacy guarantees. The company’s argument is that by building the whole stack – hardware, OS, assistant, and now cloud – it can deliver powerful AI features without exposing user data more widely than necessary. The DMA, on the other hand, is about breaking down the walls around such vertically integrated stacks, especially when they become the de facto gateway to whole markets of apps and services.

That tension is not going away. If Apple yields too much control over the assistant layer, it risks undermining its core privacy and security narrative and opening itself to unpredictable behavior by third-party systems riding on its platform. If it refuses to budge, it risks not only losing marquee features in key markets, but also further regulatory scrutiny and potential fines for non-compliance. For the EU, meanwhile, bending too far to Apple’s architecture would set a precedent that other gatekeepers could cite when they want to bolt on their own proprietary intermediaries to blunt the DMA’s effect.

From a user’s point of view, especially in Europe, the question is simpler: will I eventually get Siri AI on my iPhone, and if so, at what cost to my privacy, choice of assistants, or both? Apple says it “hopes” to bring Siri AI to the EU and will continue engaging with regulators, but the absence of even a vague window – “not this year,” “not before 2027,” anything – suggests the company is prepared for a long haul. If we have learned anything from earlier DMA battles, it is that interim solutions can look very different from what either side originally envisioned.

It is also a reminder that the AI race is not just about model sizes, benchmarks, or slick demos at developer conferences. It is about how those models actually reach users, who gets to mediate their access, and which jurisdictions are willing to draw bright lines about data access and control. The Siri AI delay in the EU is one of the clearest signs yet that in this era, the new killer feature on your phone may arrive late – or not at all – not because the technology is not ready, but because the rules for deploying it are still being fought over in Brussels, Washington, and beyond.