What My Audience Really Thinks About AI on Smartphones

We may earn compensation from reviewed products. Learn about our editorial policies.

I’ve mentioned this before, but I feel I need to illustrate the point once again: I’m not against AI. At all.

There’s a very simple reason for this – it has transformed operations behind the scenes in my business. Tasks that once took thirty minutes now take seconds. Our YouTube video titles are better because of AI and, as a result, often perform better. The need to spend thousands on sound treatment for my studio was erased, overnight, thanks to AI.

I’m just desperately bored of consumer-facing AI.

Don’t get me wrong – I’m impressed by it. It’s hard not to be when you can make stuff vanish from photos and tweak the individual audio components of a video shot on a smartphone.

The problem is that smartphone manufacturers appear to have fallen on AI as a clutch during their launch campaigns for new devices and treat it as a giant pissing contest against the competition. 

Clutches and pissing contests in tech are deeply uninteresting.

Last week, following the launch event for Samsung’s S25-series smartphones, I decided to ask my YouTube audience for their opinion on AI. The answers were fascinating, but first, I’d like to go on a bit of a preamble.

Unfinished vs polished smartphone AI

There are two types of smartphone AI to which we have access at the moment: the unfinished and the polished.

Let’s start with the latter, which is top of mind for me following Samsung Unpacked last week. Samsung’s Galaxy AI is level-pegging with Google’s AI platform. They are both equally as impressive, predictable(ish), and about as non-beta as one might hope for in 2025.

Then, we have the unfinished AI. In fairness, all tech is unfinished these days; how often do we read the footnote ‘coming later this year’ applied to new devices and software releases? But unfinished AI is about as comical and pointless as they come.

Here’s looking at you, Apple. I’ve made my thoughts on Apple AI abundantly clear already in my Apple intelligence review, so I won’t regurgitate all of those thoughts today. Suffice to say, I think Apple Intelligence is just a bit crap. Rushed. Poorly implemented. Hard to discover, let alone use. Confusing. And lots more horrible phrases that the Apple Intelligence team will not want to read.

I should confirm at this juncture that I’m a huge Apple fanboy. I get blasted for being one all the time on YouTube. I’m used to it. I’m also used to Apple sometimes being unbearably frustrating when it comes to their approach to certain products (AirPods Max ‘v2’, anyone?). Apple Intelligence, however, is an anomaly, because it feels like they didn’t want to build it. Although we’ll never know the truth – and I certainly have no evidence to suggest it’s the case – I firmly believe that, given the choice, Tim Cook and his team of execs would have voted to skip the AI revolution altogether. The future, as far as they’re concerned, is visionOS 2 – not the ability to rewrite a text message to make it sound more ‘friendly’.

I still have no idea what that means or why you’d need it, by the way.

The disparity between Samsung and Google’s AI platforms and Apple Intelligence is comical. The latter is awful at removing things from photos and turns to ChatGPT for the simplest of tasks. Worse still, it hasn’t done anything to radically improve Siri, which remains the worst AI helper on the planet. Sure, Apple Intelligence is in beta, but it was marketed with such ferocity during 2024 that we were all understandably expecting a lot more.

What I think about smartphone AI

Although it might sound like it, I’m not against AI on smartphones at all. I just don’t want it shoved in my face. It looks desperate on behalf of the brand and suggests there’s nothing left to be done with these devices beyond turning them into needless assistants for everyday tasks with which we never asked for help.

However. The side of smartphone AI I can get on board with is the stuff that’s been going on behind the scenes for a lot longer than we could remove an unwanted person from the background of a photo.

This kind of AI (or machine learning, to which it was once referred), is felt most keenly when we take photos with our phones. Anyone who has got into digital photography with a proper camera will know how frustrating it can be to grab a perfect photo – even if you spend hours faffing about with it in Lightroom after the event. This is because it relies too heavily on the skills and knowledge of the person operating the camera. By contrast, if you use your smartphone to take a photo anywhere and in whatever the prevailing conditions happen to be, it will deliver the perfect image – immediately. No tweaking of camera settings, no shooting in RAW, no editing in Lightroom – just one tap of the shutter button.

Smartphones can do this because they take countless frames in one go and use AI to pick the best elements from each one, while simultaneously adjusting individual elements of the finished photo to ensure it’s sharp, detailed, well-balanced, and properly exposed. And they do all of that in a fraction of a second – billions of calculations hidden behind that innocuous white shutter button.

That kind of AI is life-changing; it means you never miss a shot and, more importantly, you don’t need to know what’s going on behind the scenes. You just know it works every single time.

AI has indeed been helping us with telephone calls for a while now, too, by ensuring that the person on the other end of the phone can hear you, regardless of whatever carnage you happen to be surrounded by at the time. Our phones have also become pretty adept at guessing what app we might want to open, and where our next destination is likely to be, based on our behaviour.

I want all of this stuff. You almost certainly do, too. But AI features which provide initial ‘wow’ moments during stadium-sized phone launches rarely make it into the national consciousness. Because no one asked for them.

What my audience says

This is the result of my smartphone AI poll on YouTube:

As you can see, most people find AI only marginally useful on their smartphones. The smallest percentage of the 1,500+ people who engaged with the poll called it either ‘very important’ or ‘important’.

Dig into the comments section of the poll, however, and there’s some interesting feedback for the brands who are wielding this stuff.

“If I can finish a phone call and ask AI “take me to Dave’s office and remind me when I arrive to post my letter” and it recognises Dave from my contacts, fetches his work address and then notifies me to post my letter, that is the sort of thing I’m waiting for.”

“I pretty much replaced Google with ChatGPT/Gemini these days, and I love AI-driven image editing features.”

“I use AI for some assistance in scripting occasionally. But there’s still a lot that it gets wrong.”

“I still judge a phone by what it can do in my daily life, like eco system, general features that make my phone use easier, nice hardware, good battery life, snappy operating system, good connectability <sic> and safety features. Some of the AI stuff is pretty gimmicky, I wouldn’t base my purchase on AI alone.”

“It’s very important because it’s going to be running our phones for us soon. I don’t think most people understand this yet.”

“It’s a nice to have but I feel like realistically most people won’t use it all that much. Wish they would focus on other things.”

“Like 80% of smartphone features, it’s cool, and a great demonstration of technology. But it’s not something I ever use in the real world. Maybe 2/3 times a year. Absolutely hate AI art.”

“When telephone first became mobile, it wasn’t hugely impressive either. This is no different. This is the worst it will ever be.”

Interesting, right? It’s a mixture of hope for the future, caution, and acceptance that, as the last commenter noted, this is the worst type of AI we’ll ever experience.

So, what DO the people want?

Reviewing smartphones is getting harder with each new release. They’re all very capable in terms of performance, they all have great battery life, there isn’t a single display that doesn’t impress with its brightness and sharpness, and both iOS and Android are as polished and reliable as they’ve ever been.

This is why, in 2025, I’ve decided to simplify my smartphone reviews. It started with a desire to focus primarily on whatever the brand deemed was the standout feature of their new device. However, just four weeks into 2025, it’s clear that they’re all going to simply say “AI <insert something else here>”.

So, that ain’t gonna work.

Then, I had a brainwave, thanks to Samsung Unpacked. That event was so tightly packed (pun not intended) with AI, that it immediately became clear that every review would essentially be a regurgitation of the new (and old) robot-driven features, along with a brief opinion on each one. I’ve been there, done that, and got the t-shirt, and I can confirm that it’s pretty soul-destroying. And a metric shitload of work.

What if, I pondered, I just reviewed the S25 Ultra as a phone? You know – the ultimate personal device that we all rely on every single day to communicate, entertain ourselves, and get stuff done. If AI enters any of that stuff in any useful way, I’ll cover it. If not, I won’t.

Decision made. That’s what I’m going to do this year, and I’m confident it’s exactly what the people (and by that, I mean, specifically, my audience) want.

At least, I hope that’s the case. I guess we’ll find out.

Final thought

Artificial intelligence is going nowhere. And by that, I don’t mean it’s destined for obscurity. On the contrary, it will continue to infiltrate our lives, whether we like it or not. Jobs will be lost because of it, and new careers will be forged because of it.

For smartphone AI, things need to develop significantly before I stop writing opinion pieces like this. We need to reach a point where our phones become digital companions that understand our intent. That’s where the magic lies – AI which is completely transparent and so useful we forget it’s even there. Several smartphone manufacturers would have you believe that we’ve reached that point already, but I think the journey is going to be a long one.

Oh, if you’re wondering, none of this was written with AI. I’m old-fashioned like that.