Today’s most hyped AI apps don’t do much for me. Indiscriminate content scraping soured me on image generation, and I avoid large language models when possible. The underlying technology fascinates me, but I don’t seek out problems for it to solve. Still, we can’t avoid it. Smartphone sales have plateaued for years, so manufacturers need to keep consumers interested. Galaxy AI’s attention-grabbing features often look impressive and can do remarkable things. I’m not here to disparage AI in general.
AI’s processing demands justify new devices that exceed the average consumer’s needs. For example, the high-end Galaxy Tab S10+ hardly outperforms the Tab S9 FE+ in streaming movies and browsing social media. However, only the nearly $1,000 S10+ supports Samsung’s nascent toolkit. I argue that Galaxy AI isn’t useful enough to justify an expensive, new toy that tackles everyday tasks a bit quicker than its half-price siblings.
In the AI era, Google, Samsung, and Apple have made us all beta testers
Say goodbye to finished software
Galaxy AI features are still new and nebulous
Great on paper, less so in real life
Sure, Adobe Firefly, that’s how humans usually sit.
Experts and users generally agree that Galaxy AI offers less variety and utility than Google’s offerings. While AI feature criticisms apply to multiple platforms, Samsung’s practically carbon-copied new Galaxy phones fall in the crosshairs.
Tools like live conversation translation and post-recording slow-motion video conversion show promise. They can make life better for everyday folks with ordinary mobile computing habits. With a few taps, they can connect us with neighbors or dramatize recordings of our dogs soaring for a ball. That is, if they work.
Some AI features work great, and others, not so much. Consistency, or lack thereof, is the biggest complaint about today’s much-talked-about software tricks. If an app has trouble transcribing a conversation and then struggles to translate it, what’s it worth? Magically editing awkward limbs or removing strangers from a photo’s foreground makes life easy until it still looks weird after six tries. With so much potential controversy surrounding generative AI, the gatekeeping of image creation and social media proofreading sometimes breaks features.
‘AI-powered’ remains a somewhat arbitrary distinction
As muddy and misleading a definition as ever
Stroll down social media lane, or interview some smartphone-savvy circles, about the most popular AI feature. One answer leads by an incredible margin: Circle to Search. A slide of the finger produces a link to whatever you circled. The perfect example of AI’s usefulness, right?
A screenshot snipping tool tied to image recognition and fed to a search engine doesn’t encapsulate the impressive technology we call “AI.” Google Lens is seven years old, and Circle to Search is the evolution of 2015’s Now on Tap. None of this is new or worthy of an upgrade to a costly Galaxy flagship.
AI has a point, and overhyped apps miss it
2024’s second Galaxy Unpacked event introduced the ability to draw a dinosaur on a beach.
AI features inspire and depend on some groundbreaking techniques. However, focusing on the flashiest, most shocking results belies what it’s often good for. We shouldn’t need Samsung to blow our minds with functions that a slim, battery-powered device has no business performing. That isn’t AI’s only contribution.
Its benefits offer considerable potential under the hood of everyday software. Circle to Search’s effectiveness and popularity stem from the unprecedented technological support behind modern image recognition and language parsing. Counterintuitively, seemingly bog-standard tools not emblazoned with the AI label could benefit even more than apparently magical AI apps.
Arm’s senior VP explains AI’s true impact, and it’s not what you think
Our interview with Arm exec Chris Bergey
Automations like Samsung Notes’ formatting, summarization, and spellchecking (some of the few regularly praised Galaxy AI-exclusive features) don’t embody what “artificial intelligence” intelligence is, but people consistently find them useful. Anything but flashy, Notes’ tricks still use the advanced parallel processing methods and data-backed training popularized by AI madness.
AI makes a less-than-obvious difference industry-wide by influencing specialized microchips, complex programming techniques, and frameworks like Arm ISAs and their AI-focused extension libraries. Automatic note formatting may be mundane, but it helps and underscores the technology’s usefulness. Then again, mundanity plays another big part.
Modern AI features aren’t transformative
They often feel tacked on, gimmicky, and superfluous
A truly epic AI image with significant compression artifacts.
When I read editor Will Sattelberg’s Galaxy Z Fold 6 review, I assumed “sunglasses on dogs levels of gimmicks” was an idiom that had passed me by. Instead, Will was talking about his foray into Samsung’s Sketch to Image. Like many, he experimented with Sketch to Image’s novelty a few times and then forgot about it.
AI is supposed to change our lives. Galaxy AI barely changes our phones. This may be blasphemy, but on-demand AI-generated wallpapers don’t add much substance to the mobile experience. Diverse wallpaper apps and catalogs provide more choices than you can wade through. We could (and might) spend an entire article roasting the banality of outsourcing text messages to a supercomputer.
Sometimes, AI features aren’t there but somewhere else
Samsung’s AI superiority claims only apply to some regions. Many countries and languages lack access to useful aspects. Potential European Galaxy buyers see this often due to legislation curbing harmful AI use. If you don’t know when you’ll get Galaxy AI, don’t pay for a phone that can use it.
On the other hand, phones’ improving performance and efficiency allow for complex on-device operations. Galaxy AI can theoretically translate conversations, record transcripts, and perfect message grammar, even without internet access.
AI on Android isn’t worth buying into just yet
If you’re looking to buy a smartphone, today’s AI tools aren’t enough of a selling point
Still, a phone can’t do everything. The cloud takes care of generative photo editing. You need to upload videos to convert them to slow motion. Summarizing texts and recording only work online. Some translation features work better with internet access. Why fork over a month’s rent for a device that can’t organize your notes without a supercomputer’s help?
AI-powered functionality could be so much more
The most eye-catching features feel too superficial to represent how AI might improve our enjoyment and workflows. Image generation (which rumors indicate Galaxy AI will soon implement) provides a fun distraction or saves time designing social media promotions. Still, it isn’t life-changing or driving me to shell out $1,500 on the hottest new foldable.
Where’s the innovation? In an October 2024 interview, Samsung’s Head of Customer Experience mused on a future where AI adjusts settings before users need to manually. Most outlets ran with a sensationalist claim that AI would “replace the Settings menu,” a concept rightfully deserving of the widespread ridicule it drew. Nobody wants half-baked prediction algorithms reconfiguring their phone.
That isn’t far from the brazen creativity that could spur a major reimagining of mobile devices with AI’s help. The multimodal and agentic AI hysterias clogging the discourse could drive the reinvention of human-device interaction if they can mitigate privacy and human agency concerns.
The underdeveloped Rabbit R1 showed flashes of inspiration while failing to serve much purpose. Qualcomm, T-Mobile majority owner Telekom, and startup Brain AI demoed an app-eschewing phone at MWC 2024 that relied on a prototype AI agent to replace the interface as we know it. Still, those mock-ups are light years away from Galaxy AI’s fun, but not industry-shaking, set of magic tricks. We won’t reach unknown frontiers while wasting battery power by turning sketches of stovepipe hats into pictures of actual pipes.
This concept phone wants you to go app-free, but not entirely
It’s not unlike the Rabbit R1, but in a more familiar form factor
Galaxy AI isn’t done innovating yet
Samsung’s best is (hopefully) yet to come
Easily beating Samsung to the punch, Google’s world-leading data hoard and programming talent resulted in Gemini tools working consistently better than Galaxy AI. Still, Samsung won’t give up. The One UI 7 beta introduced writing tools, call transcripts, and an attempt at unifying all your available data into granular suggestions improving your daily life, such as, “Take a nap, then pack for your trip.”
Convenient, streamlined features improve the near-future smartphone experience, at least as much as a cat wearing glasses. Meanwhile, advanced machine learning techniques push researchers toward complex goals like mapping the human brain. World-altering academic study benefits from AI’s progression, and the consumer drives much of the hype and spending that fuels the machine. If Samsung dares to be different (in its features or phones), Galaxy AI could become a selling point.
The engineers have their work cut out for them
You can’t convince me that Galaxy AI does enough to warrant replacing a two- or three-year-old flagship with the Galaxy S25. Samsung’s AI toolkit fails to meet its lofty goals reliably and needs multiple doses of real-world utility. Features like AI notification and text summary, slated to arrive with the S25 family, look promising but won’t upend the industry.
Samsung can learn from Google as they co-develop AR wearables that might leverage the Project Astra AI-powered assistant. Upcoming LLM additions and agentic AI upgrades to Bixby could turn the long-suffering voice assistant into a legitimately helpful pocket-size personality. Samsung makes great Galaxy phones, but it needs to get creative and use Galaxy AI for something we haven’t seen before. It may take a bit of bravery, but it could pay off and extend Samsung’s long-running lead in the Android market.
