The European Commission has ordered Google to grant rival AI assistants substantially deeper access to Android by July 2027, while separately requiring the company to share anonymized Google Search data with eligible competitors starting January 2027. At the same time, Apple announced it will not launch its new Siri AI on iPhones and iPads in the European Union, citing privacy and security risks from interoperability mandates. These parallel moves mark the biggest regulatory showdown yet over who gets to be the operating layer on your phone.

What the EU Ordered

Under the Digital Markets Act (DMA), the Commission specified two areas where Google must change its practices. The Android interoperability decision requires Google to give third-party AI services access to operating-system capabilities that currently only Gemini enjoys—like voice invocation, screen context, app interaction, and on-device resources. The second decision forces Google to share anonymized Search data (queries, clicks, rankings) with rival search engines and AI chatbots so they can improve their services.

The timeline: Google must start sharing Search data by January 2027. Android changes will roll out in phases, with major obligations expected by July 2027. That means within roughly 18 months, European Android users should be able to set a preferred assistant—say, ChatGPT or a privacy-focused alternative—and have it work as seamlessly as Gemini does today.

Apple’s response is more hostile. After introducing a new AI-powered Siri in June 2026, the company said it would not launch the assistant on iPhones and iPads in the EU because the DMA could force it to give rival assistants access to the same deep device context and personal data that Siri uses. The result: a patchwork where EU users get Siri AI on some Apple devices (Mac, Vision Pro) but not on the phone or tablet that actually fits in your pocket.

Why Your AI Assistant Is About to Get Competition

For the average person, the most visible change will be choice. Today, Gemini is the only assistant on Android that can truly understand what’s on your screen, act across apps, or respond to a wake phrase while the phone is locked. Others, like Alexa or ChatGPT, are mostly confined to their own apps. After the EU’s rules kick in, you could have multiple assistants that compete on smarts, speed, and privacy—not just on which company preloaded them.

But that choice comes with new complexity. You’ll likely see more consent screens asking what personal data an assistant can access: messages, screen content, location, app usage. Instead of one blanket permission, the ideal system (and what regulators envision) is granular controls—allowing an assistant to read your current screen only when you ask, not, say, scanning your email in the background.

For iPhone users, the immediate impact is negative: you simply won’t get the latest Siri features until Apple and the EU resolve their standoff. If you rely on Siri for daily tasks, you’ll be stuck with the older, less capable version while the rest of the world upgrades.

Enterprise and Windows users should pay attention too. Once Android opens up, companies like Microsoft could tie their Copilot AI more deeply into your phone, creating smoother workflows between a Windows PC and a mobile device—checking your Teams calendar, attaching a file, or continuing a task across screens without Google as the middleman. But administrators will need new tools to control which assistants can access work apps and corporate data.

The Long Road to Open Platforms

This isn’t the first time Europe has forced tech giants to loosen their grip. In the early 2000s, Microsoft had to offer browser choices on Windows. More recently, the DMA itself mandated alternative app stores, payment systems, and messaging interoperability. The battle over AI assistants is the latest front, and it stems from a fundamental fear: if only the platform owner’s assistant can truly work across apps and understand your device, then switching becomes nearly impossible—locking you into Android or iOS even if a rival makes a smarter AI.

Google argues that Android already lets users switch the default assistant, and that granting deep system permissions to third parties would create security nightmares—malware masquerading as a helper that reads your messages, initiates transactions, or exfiltrates files. Apple echoes this, saying its vertical integration lets it protect users in ways that a free-for-all cannot.

Privacy experts, however, caution that these protests ring hollow when the same companies have faced repeated scandals over data handling. “I would say that we have to take these tech companies’ arguments with a grain of salt,” Calli Schroeder of the Electronic Privacy and Information Center told CNN.

The truth lies somewhere in between. Letting any assistant wield the power to see your screen and act on your behalf is risky, and the industry still hasn’t solved problems like prompt injection—where a hidden command in an email or website tricks the AI into doing something dangerous. But regulators believe the solution is to mandate secure, auditable interfaces, not to hand the keys only to Google and Apple.

Steps to Take Right Now

If you live in the EU, here’s what you can do today and in the coming months to prepare:

  • Check your current assistant settings. On Android, go to Settings > Apps > Default apps > Digital assistant app. Note which service is active and explore alternatives from the Play Store to see what’s already available.
  • Understand permission hygiene. When the new assistants arrive, avoid tapping “allow all” out of convenience. Ask yourself: Does this assistant really need access to my location, messages, or screen? Turn off background access unless necessary.
  • Read privacy policies. Some assistants process everything on-device (more private, less powerful), while others send data to the cloud. Look for clear statements about how your information is used and retained.
  • For parents and employers: Start a conversation now about which AI tools you’ll permit on managed devices. In managed environments, you may soon need to set policies that block certain assistants or limit what corporate data they can touch.
  • Stay updated on rollouts. Google will likely announce API details and developer guidelines later this year. Apple’s stance may shift if it negotiates a compromise. Follow official Google and EU channels for notices.

What Happens Next: 2027 and Beyond

The compliance deadlines—January 2027 for Search data, July 2027 for Android—give tech companies breathing room, but also create a window for legal challenges. Google has already called the rules “the wrong track” (Android president Sameer Samat on X). Court battles could delay implementation, though the DMA’s design is meant to be binding until overturned.

The real test will be whether the technical details match the ambition. If the APIs Google provides are so restricted that rivals can’t match Gemini’s performance, the mandate will look like a hollow victory. Conversely, if open access leads to a major privacy breach, public trust in both regulation and AI could crater.

For Apple, the next move is likely a negotiated settlement. The company rarely abandons a major market indefinitely. Expect a proposal for an on-device “broker” system that lets Siri AI talk to third-party services through a secure intermediary, granting enough functionality to satisfy regulators without exposing raw data. Such a system could take years to build, meaning EU iPhone users may wait past 2028 for Siri AI.

In the near term, rivals like OpenAI and Microsoft stand to gain the most. OpenAI’s ChatGPT already runs on 30% of EU smartphones, according to Omdia, and the new rules could let it finally break out of its app sandbox. That disruption—a chatbot replacing parts of Android’s interface—is precisely what Google fears and what Europe hopes will spur genuine innovation.

The bottom line: the EU is drawing a line in the sand. It’s saying that the AI assistant you use shouldn’t be determined by the phone you bought. Whether that ideal translates into safer, fairer software or a messy experiment in forced sharing will depend on the engineering—and the enforcement—that unfolds over the next two years.