For as long as the mobile phone (and now smartphone) market has existed, it has used some sort of race of numbers and metrics. Twenty years ago, it was the number of colours supported by screens (colour screens started with 4 colours) and the number of "voices" in polyphonic melodies. Then there were various races - for the thinnest body and the smallest size (it all ended quickly, because they reached the limit of convenience). For the number of megapixels in the cameras (this lasted quite a long time and hasn't stopped in some form even now). For the resolution of the screens (after FullHD, it's hard to see the difference, and QHD only makes sense for cardboards, but they never caught on). And for the number of gigahertz in the processors or cores in them (now it's not so relevant anymore). What will the race be about in the near future? The answer was given by Samsung a year ago, when it announced the Galaxy AI generation of smartphones, which began with the Galaxy S24. And with the advent of the Galaxy S25 series, we have already received the second (or third, if we take into account the updates for Fold 6/Flip6) generation of "smartphones with artificial intelligence". Therefore, over the next few years, the entire smartphone world will revolve around AI, and the metrics will probably be neuroprocessor cores (NPUs) and their computing capabilities.
Accordingly, all the new features that were announced today in Galaxy AI will gradually appear in previous Samsung smartphone models, and then they will gradually penetrate lower-level smartphones until they reach budget phones (it will take a few years, I think). Obviously, this will be closely related to the availability of NPUs in manufacturers' chipsets, which will determine the ability of smartphones to use AI in their work. At the same time, Google will develop its Gemini, which will also move towards the penetration of AI in smartphones. And so we will live in the midst of this "AI in every iron" festival until some new idea emerges to replace the focus on artificial intelligence. However, the smartphone industry has been searching for a new "big idea" for 5-7 years now, and it is not visible on the horizon, and AI is unlikely to be it, because it is a general trend in technology that is really changing the world here and now. And smartphones are just a fragment of this mosaic.
Two AI engines: Gemini and Bixby

Galaxy AI settings in Galaxy S25 smartphones. Screenshots: gagadget
As I said before, in addition to Samsung's own achievements in the field of artificial intelligence, Google is promoting its own AI called Gemini (or should I say Android or Alphabet in general? things have become very complicated). In fact, it is not for nothing that both companies started working together on the development of the platform, because who else but Samsung is currently creating reference Android smartphones? Therefore, it is the new Galaxy that receives all the preferences - both from the operating system developer and the chipset manufacturer - the Galaxy S25 also uses special versions of the Snapdragon 8 Elite processor, which have "for Galaxy" in its name. It's hard to say what the difference is between the regular Snapdragon 8 Elite - the official information refers to "optimisation" and "integration" with Samsung software. But from the user's point of view, you can expect something like overclocked capabilities (although it seems that we are not talking about overclocking, that is, the increased frequency of the processor, but rather the effect).
That's why artificial intelligence is present in Galaxy S25 smartphones in two guises: Gemini from Android and Bixby from Samsung. The former will be responsible for the "outside" world - the browser, search (Circle to Search has become even more accurate and functional), YouTube, etc. And the Bixby AI capabilities are used to control the smartphone's applications (Calendar, Clock with alarm and timer, Notes, Messages, and Caller ID). It is also responsible for Samsung's proprietary AI capabilities that appeared in the last generation of Galaxy AI (Call Assist, Writing Assist, Note Assist) and the new ones we saw in 2025 (Sound Control, Photo Screensaver).
Of course, when we give some data about ourselves to a cloud service, the question of privacy inevitably arises. Samsung assures us that everything is fine here - Knox Vault is responsible for the security of information, using "post-quantum cryptography", whatever that means.
Native Android AI: voice control
In 2025, thanks in no small part to a new processor, smartphone AI will become multimodal. This means that artificial intelligence can work with voice, text, images, and video. That is, you can open a three-hour video interview on YouTube and extract any information that a large language model understands from it in seconds. For example, how many questions were asked during the interview, which question took the longest to answer, and how many times a certain word was used. This way, you can quickly find the desired fragment or simply ask the AI to make a summary of the entire video - what was discussed there. Other voice-related features can be added to this list: reading text aloud and translating it, and the ability to make (and read aloud) a summary of a large text. I wonder what, for example, will remain of this text of mine if I ask AI to make a summary of it.
The important issue of supporting the Ukrainian language has already been implemented here, as Gemini already works with Ukrainian voice. Interestingly, it stores the entire history of your conversations (it's like ChatGPT with one permanent session). Therefore, you can assume that you are constantly training the AI, and the more you talk and tell about yourself and your preferences, the better it understands them and offers relevant answers or actions. The power of artificial intelligence lies in its ability to process large amounts of data, which forms the basis of these "predictions" of user actions. So, the main news for us is that voice search is becoming multimodal, and this improvement also applies to Circle to Search. It will also be possible to change the settings of the smartphone itself simply by voice. And this will work not for individual voice commands, but in natural language - as you are used to speaking.
However, the expected ability to translate calls into Ukrainian has not yet been implemented. Unofficially, Samsung does not comment on why. From unofficial conversations, we can understand a rather unfortunate thing: the headquarters does not understand the feasibility of investing in the Ukrainian version of artificial intelligence for translation, since most Samsung smartphone users in Ukraine have Russian in their system settings. Of course, this applies not only to new smartphones, but to all the company's smartphones in the hands of Ukrainians. It should be noted here that, like all normal businesses, Samsung uses numbers and facts to make business decisions. And they are notoriously stubborn. In general, it is a shame, but this is the situation so far.
Personal concierge: Now Brief and Now Bar
Apart from new AI assistants (discussed below), the main features of Galaxy AI 2025 are two new AI tools: Now Brief and Now Bar. The former is a welcome screen that wakes up after unlocking the screen. It provides some current information (for a newly unpacked smartphone, it is a greeting, weather forecast for the day, and calendar entries for today). The content for this screen can be managed (in the screenshots below, the list is actually a little bigger than in the third image). Now Brief has two parts: the morning report and the day's summary. In the morning, this screen will tell you about your sleep status, weather, and event schedule. It will also provide suggestions for operating modes that match the scenarios you use your smartphone for: if you drive to work every morning, it will switch to the Driving mode at a certain time and launch your favourite music streaming service (if you listen to it in the car, of course).
At the end of the day, the panel will summarise the day, show you the photos you took during the day, and remind you of upcoming events (this can include not only calendar entries, but also reminders, for example, about cinema tickets you have bought.

The Now Brief panel and its settings. Screenshots: gagadget
The Now Bar is an interactive notification window that appears on the locked screen of a smartphone and allows you to perform certain actions with applications without launching them. How this will work in general (and with which applications) is not yet very clear. But Samsung's general concept in this regard is to provide "personalised AI", whatever that means now. But it is quite clear that we are on the verge of creating what the company calls a personal concierge. That is, a kind of Jarvis that Iron Man had - a voice assistant that knows all your habits and preferences and provides advice or reminders at any time. Even when you don't directly ask it to do so. In general, this is more about the future than about today. After all, the value of any system (and an AI assistant is no exception) increases with the number of users and applications. There is a classic chicken-and-egg trap here: until there are users, developers will not adapt their software and vice versa. But with AI, sooner or later, humanity will inevitably come to personal AI assistants. It is only a matter of time. This story is being created right before our eyes, and everything is happening right now.

Setting up the Now Bar. Screenshots: gagadget
Other AI features: sound control
Last year, we introduced AI capabilities for photo editing, and now it's time for video editing. And this wonderful era begins with the simplest thing - editing sound in video. And it's not just social media users who post their vertical videos that will appreciate these features. Now, you can edit sound quickly and relatively efficiently right on your smartphone. AI can distinguish between six types of sound that can be affected (each using a separate neural network, or even more than one): music, voice, wind sounds, nature sounds, crowd sounds, and normal background noise. And each of these parameters can be reduced or increased from 1 to 100 points. It's also interesting that you can use end-to-end video recording from your smartphone when recording video. That is, start the desired music track on your smartphone and start recording video. The sound will be recorded from the selected song. A small thing, but interesting for social media. Now it will be even easier to make an attractive video without additional editing.

Sound control settings. Screenshots: gagadget
What else is new in Galaxy S25 smartphones
Of course, in addition to artificial intelligence, the new Galaxy S25 smartphones have several new features. Of course, most of them look predictable, we see this every year from the reduction of bezels (they are getting smaller, but they will not disappear forever, although everyone promises "bezel-less screens" for many years), increased processor performance, etc. Briefly about the main thing:
- Snapdragon 8 Elite processors for Galaxy - performance compared to Galaxy S24 has increased by 40% for NPU, 37% for CPU and 30% for graphics
- Performance with ray tracing in 3DMark increased by 40%, FPS in Vulcan - by 18%
- Steam camera increased by 40% in Galaxy S25 Ultra and 15% in Galaxy S25 and Galaxy S25+
- Galaxy S25 Ultra camera has a 200 MP + 50 MP + 50 MP + 5x optical zoom setup
- Night video recording has been improved with an SPF (Spatial-Temporal Filter) that recognises moving and stationary objects
- HDR has become 10-bit instead of 8-bit (this made it possible to increase the dynamic range by 4 times)
- Changes in design - thinner bezel (I really liked the look), less weight
Here are a few more photos comparing Galaxy S25 Ultra with Galaxy S24 Ultra, instead of a strict design, the corners are now "licked", but the sides, on the contrary, have become straight, not rounded:






In the dry balance
Of course, the phrase "a new era of artificial intelligence" in smartphones sounds a bit of an exaggeration. As we've seen for several years now, we're still seeing evolutionary rather than revolutionary changes in the new Galaxy S. Some of the new AI capabilities are derived directly from the new Android. Some of them are Samsung's own development. There are a lot of new features, but each one is not a killer. However, in combination, they all look interesting and there is simply no alternative to the new Samsung smartphones on the market. The rest are either Chinese manufacturers or iPhones (from another universe). Of course, some part of the audience may be interested in the Pixel, but its impact on the global market is not even minimal, but homeopathic. And thanks to its bare-bones Android, it remains an attractive smartphone for developers and geeks. However, let me remind you that the historical process of introducing artificial intelligence is irreversible (and smartphones are just a special case), and Samsung naturally wants to lead this trend. And, in my opinion, the company is quite successful.
For those who want to know more
- What is Galaxy AI on the example of Samsung Galaxy S24 Ultra
- AI at its best: Samsung Galaxy S24 Ultra review
- 7 useful scenarios for applying artificial intelligence in smartphones that do your work for you
- Samsung at CES 2025: what is Vision AI and how it will change TVs and monitors