CALIFORNIA, US – The Samsung S24 was a milestone for the company as the phone was the first to attempt a comprehensive package built around artificial intelligence.
Among other things, the S24 phones were the company’s first that could translate languages, transcribe audio recordings, remove unwanted elements, or extend the background in a photo through what’s called generative edit, add frames in a video to simulate slow motion, and cartoonize your selfies.
The push for AI proved beneficial for Samsung, driving sales, with about 24% of buyers citing AI as a key reason for choosing the S24, according to Kantar. In Q1 2024, sales of the S24 Ultra were about 8% higher globally than its predecessor, the S23 Ultra, with peaks of 28% higher sales in markets such as Western Europe (France, Germany, and the UK).
(READ: Samsung Galaxy S24 Ultra review: A good year to upgrade)
Samsung’s new S25, unveiled on January 22, certainly has a lot to live up to. And this year, it’s once again relying on AI to keep that momentum.
What differentiates the S24’s AI package though from the S25’s is something called “AI agent.” Don’t be intimidated. An AI agent is simply AI that can perform tasks for you such as book airline and hotel tickets, or help plan a meal, as explained by The Conversation:
“Simple AI agents can be taught to reply to standard questions sent over email. More advanced ones can book airline and hotel tickets for transcontinental business trips. Google recently demonstrated Project Mariner to reporters, a browser extension for Chrome that can reason about the text and images on your screen.
In the demonstration, the agent helped plan a meal by adding items to a shopping cart on a grocery chain’s website, even finding substitutes when certain ingredients were not available. A person still needs to be involved to finalize the purchase, but the agent can be instructed to take all of the necessary steps up to that point.”
Samsung’s S25 series is an attempt to put together something with these AI agent-like functions. Let’s discuss some of these.
Experiencing Gemini Live
First and foremost is the phone’s Gemini Live feature. Simply long-press the bottom side button (or what traditionally has been the power button below the volume rockers), and you get access to Gemini Live.
Gemini Live was unveiled in May 2024, before debuting in Google Pixel 9 phones in August 2024, replacing Google Assistant. It was praised then by outlets such as the Wall Street Journal who also found that it “sounds so human” that they forgot it was a bot.
And I agree.
We’ve had a short time with Gemini Live, but it’s been a feature that’s been kind of exciting to use because the virtual assistant may be the most natural-sounding I’ve used. It really feels like you’re talking to a person. The intonation sounds so natural, and when you interrupt it in mid-sentence, it seamlessly detects that you’re attempting to say something new, so it stops, and listens, and then responds to your newer command or question.
One other thing that I was impressed with was, during our test conversation, I asked for the cheapest tickets possible in an NBA game. And it did find tickets for those, but then it suggested that those “nosebleed” (its word) seats might not get me the best experience, and asked if I wanted a spot that was more “central.” So I agreed, and it found me what it believed to be better seats nearer to the court’s “baseline” that weren’t exorbitantly expensive ($48).
It was just a short conversation, and this is something that I seek to test further, but from this experience, it felt like it was genuinely trying to help you.
My experience with Google’s Gemini Live and its integration with the S25 reminds me of what Forbes wrote about “agentic AI systems” recently:
“Rather than simply issuing commands and receiving outputs, we’ll develop more collaborative relationships with AI systems that can engage in genuine back-and-forth dialogue, propose alternative solutions, and even challenge our assumptions when appropriate. This evolution could lead to unprecedented levels of human-machine synergy, where AI becomes less of a tool and more of a partner in problem-solving and innovation.”
I asked the S25 if it has the ability to book tickets for me. It said it didn’t have the ability to do that “yet,” and so it instead offered an alternative solution, which is to lead me to the website where I could get the tickets.
Gemini Live couldn’t yet set a schedule for me on my calendar, but the AI also explained to me that it’s something that it should be able to do in a “future update.” Samsung talked about “cross-app action” that allows its AI to use Google and Samsung apps, as well as Spotify and WhatsApp, so certainly, the phone AI’s ability to perform more functions on behalf of the user should be in the cards.
One thing that Gemini Live could already do is compose a message, and access the map app. In a demonstration, I saw how a staff asked the AI to look for a certain restaurant and send the details to a friend. The AI did just that, and composed a message. The action of sending the message is left to the human user.
Gemini Live was also able to take a photo of the contents of a refrigerator, and assess and read out a recipe for the user.
In my own test, I was also able to take a photo of a French restaurant menu, and I asked it to make order suggestions based on my budget. It did that too. And I even asked it to translate the menu items in English, and then make the order to the waiter in French.
Beyond these things, I could imagine Gemini Live as something you could just talk to when you’re bored, kind of like chatting with a much, much more advanced Amazon Echo — and one that you can have in your pocket anytime.
AI Select, natural language search
One other new function is a tool called AI Select. Once enabled (you can summon it via a sidebar), the tool lets you tap or select whatever is on the screen, and then offers tools that apply to what you selected.
For example, when you’re on YouTube, watching a sports highlight, it can offer you the GIF-making tool and make a GIF out of the video. In another case, you can take a photo of, say, a ticket, and it will offer to put the ticket details (date, event, address) onto your calendar. Lastly, you can take a photo of an object — let’s say a statue of a horse — and it can transform the object into a drawing in various styles.
The phone also touted natural language search, wherein you can do a voice or text search for things in your phone. For example, you can ask it to look for a “photo of friends eating pizza in a bar” in your gallery or you can ask it to show “settings for adjusting font size on my phone.”
Now, the S25 boasts a claimed 40% improvement in NPU (neural processing unit for AI tasks) performance, powered by the Qualcomm Snapdragon 8 Elite across all models. That means, it should be faster than the S24 when it does on-device AI tasks such as the aforementioned generative edits or language translation.
But more than the performance improvement, the S25 is banking on pushing for a role as a “personal concierge,” headlined by Gemini Live. It’s looking to build on those AI functions it introduced last year, this time focusing on making the phone easier to use with its virtual assistant functions.
Gemini Live is available on all S25 variants. The base starts at P51,990, the Plus at P67,990, and the Ultra at P84,990. – Rappler.com