If you picked up an Android phone recently and noticed a different voice answering your questions, you are not imagining things. Google has begun replacing Google Assistant with its Gemini AI model as the default assistant across Android devices, a shift the company confirmed through updated Android support documentation in the spring of 2026. The move puts Gemini in front of a user base that numbers in the billions, and it is happening months before Apple is expected to ship its own reimagined Siri with iOS 27 this fall.
The timing is deliberate. By embedding Gemini at the operating-system level now, Google is racing to make its AI the habit people reach for every day, well before Apple finishes building a Siri that can tap into third-party AI services beyond ChatGPT.
What Google has actually shipped
Google’s Pixel 9 phones have run Gemini as the default assistant since their launch in 2024, but the company has since expanded that default to additional Android hardware. Google’s support pages state that “some Android devices” now ship with Gemini out of the box, and that any user on a compatible phone can switch from Google Assistant to Gemini manually. Once the switch is made, Gemini handles voice commands, lock-screen actions, timers, and media playback.
TechCrunch reported that Google pointed to improvements in music controls, alarms, and lock-screen interactions as reasons for broadening the rollout. Those are real quality-of-life features, but they also represent a narrow slice of what Google Assistant has handled for years: smart-home routines, third-party app integrations, multi-step automations, and accessibility commands. Google has not published a detailed feature-parity list, and independent testing of Gemini’s coverage across those categories remains limited as of June 2026.
The company’s own messaging adds some confusion. A Google blog post describes the transition as replacing Assistant with Gemini on Android phones, yet also says Assistant access will be “reduced on most mobile devices over the coming months.” A separate statement references “reduced Assistant access later this year.” Whether users face a hard cutoff or a gradual feature migration is still unclear, and the two timelines do not quite match.
What Apple is building with iOS 27
Apple is taking a fundamentally different approach. Bloomberg reported in March 2026 that Apple plans to open Siri to rival AI assistants beyond its existing ChatGPT integration when iOS 27 arrives. A follow-up Bloomberg newsletter in April spotted a revamped Siri interface hidden inside WWDC teaser artwork, reinforcing that a visual and functional overhaul is on the way.
If Apple follows through, Siri would act less like a single assistant and more like a broker, routing requests to whichever AI provider the user prefers or whichever handles a given task best. That model could appeal to people uncomfortable handing every query to one company’s AI. But Bloomberg’s reporting describes intentions, not a finished product. Apple has not confirmed which third-party AI services will be available at launch, how deeply they will integrate with on-device intelligence, or whether users will be able to switch providers in real time versus picking a single default.
The expected timeline is fall 2026, consistent with Apple’s annual iOS release cadence. Until then, iPhone owners remain on the current Siri with its existing ChatGPT integration, and nothing changes on their devices.
Two very different bets on how AI assistants should work
Google is centralizing. One AI model, trained on Google’s data and infrastructure, becomes the single front door for everything you do on your phone: searching, messaging, controlling smart-home devices, playing music. Every interaction feeds back into Google’s ecosystem and, presumably, its ability to surface Google services. The upside for users is a more capable, more conversational assistant. The risk is that anyone who built routines around Google Assistant’s specific behaviors may find those routines broken or degraded during the transition.
Apple is decentralizing, or at least signaling that it wants to. By letting multiple AI services plug into Siri, Apple can position itself as the neutral layer that protects privacy and controls the interface while letting specialized models handle the heavy lifting. The upside is choice. The risk is complexity: users may not understand which AI is answering them, who holds their data, or who is accountable when a response goes wrong.
Neither approach has been tested at full scale. Google is further along in deployment but has not proven Gemini can match every Assistant feature that users depend on. Apple has not shipped anything yet and is still in the building phase.
What Android and iPhone users should actually do right now
For Android owners, the practical advice is simple: if Gemini is now the default on your phone, test it against the tasks you use most. Set your usual alarms, run your smart-home routines, try your favorite voice commands. Google’s support pages confirm that switching back to Google Assistant is still possible on compatible devices, but the company’s stated direction points toward reduced Assistant availability over the coming months. The window to compare the two side by side will not stay open indefinitely.
For iPhone owners, there is no action to take today. The fall release of iOS 27 will bring changes, but until Apple details which AI services Siri will support and how the new interface works, any preparation would be premature.
For everyone watching the broader AI assistant race, the real test is not which company announces more features. It is which assistant people actually trust enough to use dozens of times a day, for tasks that range from trivial to sensitive. That answer will not come from a keynote. It will come from the millions of small interactions that start the moment someone picks up their phone and says, “Hey Google” or “Hey Siri,” and decides whether the response was good enough to try again tomorrow.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
