Is hybrid AI the smartphone wave of the future? Samsung and Apple think so

A detail that may have slipped under your radar with Samsung’s new Galaxy S25 is its unusual approach to generative AI. While there’s a lot of hype around the adoption of Google Gemini, Samsung’s own Bixby assistant is still very much present. You can’t ask for it by name — but it handles tasks Gemini can’t, primarily a range of app functions. If you use Samsung Calendar, for example, Bixby will kick into gear when you want to add an event, without ever identifying itself as Bixby.

Can we expect to see more of this hybrid approach to phone assistants? I think so. It seems like a practical necessity in the short term, and possibly the best approach in the long one.

Sit down, be humble

The massive hurdles of AI

OpenAI

Frequently, the focus of the major generative AI platforms seems to be on creating a one-stop shop. All AIs have to turn to outside data at some point, but platforms like Gemini and ChatGPT are tackling a growing number of functions internally. Most people are familiar with ChatGPT’s ability to answer knowledge questions, rewrite text, or generate graphics — but it now also handles things like to-do lists, reminders, and web searches. It even performs data analysis and code completion. It doesn’t have deep hooks into phones just yet, but only because Apple and Google are the gatekeepers.

The current situation is probably untenable in the long run for several reasons. At a minimum, there are legal issues — Apple and Google are already under fire for monopolizing app distribution, never mind the implications of funneling all of a user’s activity through a single assistant that prefers first-party apps. There are also controversies over AI models, since while companies often use publicly accessible media for training, they don’t often obtain consent from creators. Upcoming court rulings could make it legally necessary to form more partnerships.

Beyond that, the scope of AI is only going to become more unwieldy, even in the hands of megacorporations. Think about how many services, features, developers, and datacenters are required to support commands like “rewrite this email to sound more professional” or “turn on Wi-Fi when I get to 607 Katmai.” Now imagine adding technologies like augmented reality and self-driving cars into the mix, and scaling that so that millions of people can use your AI simultaneously. We haven’t even seen much adoption of generative tech in smart homes, despite Amazon Alexa being around since 2014. I’m not holding my breath for Alexa’s rumored upgrade.

It’s not that crazy, then, to expect phone makers to combine multiple AI platforms on their devices, since one will inevitably be better at a given task than another — if it supports that task at all. It’s mostly a question of companies putting dreams of total control aside in favor of delivering a product that’s useful.

Related

Does smart tech really add that much to your fitness regimen?

The full answer varies from person to person, but here’s my gym-buff take.

The prospects for hybrid AI

Is a trend already underway?

Apple / Pocket-lint

Samsung isn’t the only phone maker following a hybrid strategy. You probably know the other one — it’s Apple. While Apple loves to brag that Apple Intelligence is compartmentalized, often running entirely on-device, the truth is that ChatGPT does a lot of the hard work. If you ask the latest version of Siri a complex knowledge question, it forwards that prompt to ChatGPT. The same goes if you trigger Visual Intelligence with the iPhone 16’s Camera Control button. I’m glad, since Siri has long been awful about knowledge questions, frequently asking you to tap a web link instead of actually answering.

The Galaxy S25’s launch could well mark the beginning of a trend, but we won’t know for sure until other brands jump on board. Motorola is preparing a platform called Moto AI, but that’s still in beta, and we have no idea if the tech might merge with Gemini elsewhere on Moto phones. To me, that’s mandatory for calling something “hybrid” AI — it doesn’t count if assistants are stuck in their own silos. Likewise, OnePlus is only beginning to dip its toes into custom AI with devices like the OnePlus 13.

There are limits to how far things can go, of course. Unless there’s an antitrust ruling, Android phones are inevitably going to pair something with Gemini, and iPhones will only ever have hooks into Apple Intelligence. Gemini is rumored to be the next add-on for Apple Intelligence, in fact, creating some sort of unholy union.

Hopefully, we’ll get lucky and see assistants support three or more AI platforms at some point, including ones geared for specific tasks. Certainly, I’d rather use Midjourney for image generation than Apple’s Image Playground.

Related

The iPhone 16’s Camera Control button sucks right now

It’s both half-baked and redundant for a lot of users.

The inevitable farewell

The trend won’t last forever

Whether or not a trend is getting underway, I’m not expecting it to last forever. Without legal guardrails, many companies inevitably give in to the temptation of total control. It presents a more profitable and cohesive package to customers, after all, and no one wants to be in someone else’s pocket, especially a rival. Google pays Apple to remain the default search option in Safari — but Google would drop those payments like a stone if Pixels sold as well as iPhones.

The Galaxy S25 could become a testbed. If the Gemini/Bixby hybrid works smoothly, I’d expect to see it come to other Samsung devices. If that happens, other phone makers are bound to pay attention, and might not worry so much about other companies having a finger in the pie.

Related

Do smartphone buyers actually care about camera notches in 2025?

Phone makers shouldn’t be in a rush to conceal them anymore.