How AI could make smartphones great again

But tech companies aren’t ready to give up on smartphones yet, and the reason is simple: it’s all about the money. 

Let’s take the case of the iPhone. Sales of iPhones hit $69.70 billion in Q1 2024, or 50 per cent of the company’s total revenue last quarter. That number is bigger (even though smartphone shipments remained flat and profits for hardware manufacturers continue to drop) than the revenue of many Fortune 500 companies, so you can imagine how profitable the iPhone is to Apple. Samsung also brings in billions of dollars in profits by selling its Galaxy smartphones. 

Both Samsung and Apple are not only the top smartphone sellers in the world, but they also command major cuts in global phone profits. Then there is Google, the maker of Android which powers a majority of smartphones and is the dominant force in Search on mobile, alongside its other apps like Gmail, Maps, and the Google Play app store. The entire ecosystem, which also comprises chip makers and developers, comes together and adds to the mobile space.

‘Gen AI comes to smartphones’

Motorola Edge 50 Pro is designed around AI smarts. (Image credit: Anuj Bhatia/Indian Express)

Once again, this year’s phones will look a lot like last year’s. But real innovation is what’s inside, both at the chip and software level — the AI. Tech companies are keen for consumers to focus on artificial intelligence on smartphones as the next big thing. Although AI has been powering phones for many years, this time the plan is to power everyday features that people use on their smartphones. 

Samsung’s Galaxy S24 and Google Pixel 8 already demonstrate how AI can be used to simplify complex tasks. This is made possible by AI, particularly generative artificial intelligence, which refers to artificial intelligence models trained on large troves of data that generate new content when you provide a prompt. Both phones run generative AI through mobile devices, which would allow AI chatbots and apps to run on the phone’s own hardware and software rather than be powered by cloud services in data centres like how OpenAI’s popular ChatGPT and Microsoft’s Copilot work. 

More than the display, design, or the S-Pen, the Galaxy S24 Ultra’s (review) marquee feature is “Galaxy AI.” Some of it is made by Samsung, others by Google. Many of the AI features require an internet connection, while others are performed entirely on-device, meaning those features can be used without necessarily being connected to the internet at all times.  

Perhaps the most polished AI feature on the Galaxy S24 Ultra is Circle To Search. It lets you search for anything on the screen just by circling it with your finger. For example, if you see a new Aviator-style sunglasses that your friend is wearing in an Instagram post, and you want to know more about them, all you need to do is press the home button for a second. This triggers a transparent filter that overtakes the screen, and a bunch of search results for Aviator-style shades pop up. Another feature called Live Translate, enables the Galaxy S24 to transform into a personal interpreter during a phone call by translating speech in real time. Circle to Search and Live Translate are among a handful of AI-enabled features introduced on Samsung’s Galaxy S24 series. 

Meanwhile, Google is marketing its new Pixel 8 lineup as legit “AI phones” from day one. So it’s no surprise that the Pixel 8 Pro leans heavily on new AI features. The Magic Editor, for example, lets you reframe your photos and remove objects. Audio Eraser, on the other hand, removes wind noise and other annoying sounds all using AI. The Pixel 8 and Pixel 8 Pro (review) come with a Recorder app that can automatically summarise your conversation and send you a few important bullet points. 

Motorola has also jumped into the Gen AI race and added a layer of AI to improve cameras and a built-in AI wallpaper in its latest Edge 50 Pro. Although the Gen AI implementation isn’t as aggressive as Samsung’s, the brand seems more interested in bringing AI features where consumers can see the difference. “We won’t just throw AI features into a smartphone and try to sell it based on that. It’s about what they are good for and what we are using them for. We will have our unique take on it as we always do at Motorola,” Ruben Castano, Head of Customer Experience & Design at Motorola Mobility, recently told indianexpress.com.

OnePlus’ announcement (alongside Oppo) that it will include Google-powered Gemini AI into their smartphones “later this year” signals no brand wants to be left behind in this race. OnePlus recently added an AI-based photo eraser tool for photos to its phones, so it has already started adding AI features to its devices but one step at a time. 

Now, the pressure will be on those brands who have neither the resources nor an AI strategy to cut deals with players like Google and add AI capabilities to their phones. Without AI-infused features, it will be hard to sell new phones, especially flagships, to consumers. 

But all eyes will be on Apple, as the entire industry is waiting for the Cupertino giant to reveal how it plans to integrate AI into its devices, especially the iPhone. Apple has been secretive about its AI plans, but according to industry watchers, the company is ready to unveil new updates to its operating systems that may include generative artificial intelligence capabilities. For a change, though, Apple’s catch-up in the generative AI boom could come with the help of Google or OpenAI, as the tech giant may collaborate with external partners to bring cloud-based AI features to its devices.

However, the challenge for Apple and others who see a future where the computing that enables generative AI, like ChatGPT, happens on the device, is how fast queries can be processed, even offline. This technological challenge requires reductions in the size of the large language models that power AI, as well as higher-performance mobile processors that make on-device AI possible.

Running large language models natively on smartphones versus on the cloud has its advantages. Privacy is one advantage, since all the processing power happens on the device, and you are not sending private data to far-off servers. However, the biggest advantage of on-device AI is efficiency, as running Gen AI applications and features on the cloud is expensive and costs a lot of money. This is why OpenAI charges certain fees from consumers to use a paid version of ChatGPT. 

‘Software will be key to AI smartphones’

Google’s high-end Pixel 8 Pro is a showcase AI smartphone. (Image credit: Anuj Bhatia/Indian Express)

The future of smartphones is less about the “phone(y)” stuff we use them for today, but rather about how AI and software work together to change the user interface, thus also altering the form factor. The AI will add a layer on top of the software and alter the interface depending on the form factor of the device, similar to the R1, an AI-enabled handheld system from Rabbit Inc., an AI startup. In reality, however, the red-orange, squarish device is more of a tool to experience the software inside the R1 or Rabbit’s operating system, called Rabbit OS, and the AI-powered technology. The best way to describe the R1 is as a virtual assistant that takes the form of a pocket computer. But this is just the beginning of AI-powered gadgets. 

“I think apps will coexist, but they will get integrated into the human interface. You may not even need to launch them; your assistant will retrieve the data you need,” Alex Katouzian, SVP and GM of Qualcomm’s Mobile, Compute, and XR, told indianexpress.com during the Snapdragon Summit in Hawaii last year. 

Smartphones will head in the same direction, and metaphorically, you will still be using them 10 years from now. Maybe they will change in shape or form factor, or perhaps they will continue to look the same as they do today, but the concept of the pocket computer won’t change. But what will really change is how AI will be deeply integrated into the hardware, software, and applications. You will buy a phone purely looking at how a phone interface is well adapted and integrated with generative artificial intelligence elements. Just think of how there are many Android skins out there. Some players use Stock Android, Samsung has OneUI, and Nothing offers NothingOS. Fundamentally, they are all powered by Google’s Android operating system, but they have a unique take on Android.

 The idea of a smartphone that is built around artificial intelligence and can be used as a personal interpreter or something completely different without connecting to the internet is intriguing. This new kind of technology has a chance to make smartphones exciting again—and possibly the reason to upgrade. This also opens up the idea for phone brands to increase their sources of revenue by charging a fee for certain features, such as AI transcribing services aimed at students and journalists. That might be the beginning of fully customised phones tailored for individuals, all powered by AI. The possibility of a new AI App Store on smartphones cannot be ruled out. 

Gemini comes in a few different sizes, and the smallest “Nano” size is specifically designed to run on smartphones, to enable “on-device AI.” (Image credit: Anuj Bhatia/Indian Express)

Perhaps the biggest question lingering on everyone’s mind is how these AI-powered features, like photo editing using Gen AI, for example, further blur the line between what is authentic and what is fake online. ChatGPT-maker OpenAI last month showed off a new technology that can recreate a person’s voice with just 15 seconds of recording of that person talking. Experts fear the tool will make it easier to execute scams and cons. OpenAI itself admits wide use of this tech would be too risky. Unless checks are put in place, AI-generated content that bears no relation to reality could be a real threat in the age of social media. 

But as some industry insiders believe, generative AI models aren’t convincing enough yet as they tend to make factual mistakes and can be inaccurate. The problem is how AI chatbots like Google Gemini have a “bias”. Despite this, Google is aggressively moving forward with advancements in AI. It develops its own AI models, cuts deals with phone makers and others to deploy its AI technology onto their devices, has its massive cloud network needed for AI processing, and has access to a gigantic user base.

The speed at which these tech companies are bringing their on-device AI models to smartphones may only increase the proliferation of AI in our lives, even though some call it “vague” and “AI overload.” With both Samsung and Google recently expanding newly introduced Gen AI features to older devices via software updates, it helps us understand how aggressively these brands intend to give a glimpse of what Gen AI can do on smartphones.

Google, in fact, expects its more advanced large language models to arrive on Android phones next year. The company’s Gemini Nano, its efficient model for “on-device” AI, can already run across its Pixel devices and certain Android smartphones. Maybe we are about to enter a new phase of smartphones with “AI” being at the left and centre.