Support us on Patreon
PayPalAI in your Ubuntu Desktop? Eventually, it seems. Canonical will need to tread very carefully on this one but plans are being made for it.
In a detailed post written up by Jon Seager, the technical leader and software engineer working for Canonical as VP Engineering, the post goes over their current thoughts on it and some potential plans to expand how Canonical use AI and how it might be added to Ubuntu directly.
"The bottom line is that Canonical is ramping up its use of AI tools in a focused and principled manner that favours open weight models with license terms that feel most compatible with our values, combined with open source harnesses. AI features will be landing in Ubuntu throughout the next year as we feel that they’re of sufficient maturity and quality, with a bias toward local inference by default.
AI features in Ubuntu features will come in two forms: first as a means of enhancing existing OS functionality with AI models in the background, and latterly in the form of “AI native” features and workflows for those who want them."
Jon Seager
Unlike certain other companies, Canonical don't seem to be trying to force anything here. Seager mentions how they won't be measuring people by how much they use AI, but where it's being used effectively where it can be "controlled and reviewed".
Seager at least seems fully aware of the issues AI tools can cause, like slop pull requests to various open source projects, and hindering people's ability to actually learn if you're just getting an AI bot to do things for you. That said, Seager believes "LLMs to be an excellent learning tool" but you still need to "be skeptical and not blindly trust what comes out of the machine".
As for what to expect with Ubuntu and AI, the framework they're developing is split between "explicit and implicit AI features". With implicit AI being enhancing system features like improved speech-to-text and text-to-speech as just one example. Whereas explicit AI would be more about "'agentic' workflows" noting that "Implicit AI features will improve what Ubuntu already does; explicit AI will be introduced as new features".
Plans are being made on how to "integrate agentic workflows into Ubuntu for those who want it in a way that feels tasteful, aligned with our user base and respectful of our privacy and security values". Seager believes Snap packages will help provide these features safely and securely.
So we're going to have to wait and see what type of AI fluff gets added to Ubuntu over the next year or two and beyond.
Article taken from GamingOnLinux.com.
See more from me
You can also find comments for this article on social media: 
Last edited by WorMzy on 27 Apr 2026 at 10:51 am UTC
Arehandoro 4 hours ago
And only the two first features should be active by default and opt-out with the first page of the installer asking if you want to keep them active or not. And the third features being opt-in with no model pre-uploaded (but uploading a local model being easy).
Because they are accessibility features.
ZeroPointEnergy 4 hours ago
If they go in a way that gives full control to the user, with like a checkbox/prompt before enabling something new, who cares. If they go in a way that forces things on because "it's ok, trust us bro" and "who cares, it's just a little harmless extra nail in the coffin", then no. And, well, with Canonical really liking to force things over… we'll see I guess.
I know there are alternatives out there, but the less people will care about this, the more it will become prevalent. And when every major distro decides that it's ok to do that, we're screwed.
Call it what it is, chatbots that rewrite text badly, then blame you for anything that went wrong.
(I'd really love to have e.g. a search feature through my huge picture archives. Or translation. I guess TTS and STT would be great too...)
Although I do see huge issues with how, for what and mainly by whom the AI is being used these days - and some huge negative questions and implications attached nobody has an answer for - there are some interesting and really useful usecases...
Quoting: Cat_fanIMO there are only three features using LLM which should be installed by default (on a desktop setup): LLM based text-to-speech, LLM based speech-to-text/voice commands and image description tied to tts.You say that but there is one little problem. Where should they run? If it's an online service it shouldn't and can't be on by default. And if it's run locally, it'll take quite a bit of power to do so and the "accessibility" of it becomes a lot less. I which case a more targeted and less brute force solution might be preferable.
And only the two first features should be active by default and opt-out with the first page of the installer asking if you want to keep them active or not. And the third features being opt-in with no model pre-uploaded (but uploading a local model being easy).
Because they are accessibility features.
Quoting: EhvisWell if you happen to posses more recent hardware (which also will be more are more prevalent and standard in the future) you will necessarily have either NPU or some similar possibility of hardware accelerating a LLM. In that case I believe it might not be such a big deal performance-wise...Quoting: Cat_fanIMO there are only three features using LLM which should be installed by default (on a desktop setup): LLM based text-to-speech, LLM based speech-to-text/voice commands and image description tied to tts.You say that but there is one little problem. Where should they run? If it's an online service it shouldn't and can't be on by default. And if it's run locally, it'll take quite a bit of power to do so and the "accessibility" of it becomes a lot less. I which case a more targeted and less brute force solution might be preferable.
And only the two first features should be active by default and opt-out with the first page of the installer asking if you want to keep them active or not. And the third features being opt-in with no model pre-uploaded (but uploading a local model being easy).
Because they are accessibility features.
Mountain Man 1 hour ago
Last edited by Mountain Man on 27 Apr 2026 at 2:20 pm UTC
Quoting: Mountain ManI hate AI. It is being incorporated into the ticketing system at my job, and the three paragraph AI "summary" is often confusingly written. AI seems to take an "everything plus the kitchen sink" approach that provides redundant and unnecessary information, and I can get a better handle on the issue simply by reading the customer's original brief question.It's very much like the kid who didn't study the class material but still tries to write a convincing essay and make it sound like he's smarter than he really is.
Quoting: Cat_fanIMO there are only three features using LLM which should be installed by default (on a desktop setup): LLM based text-to-speech, LLM based speech-to-text/voice commands and image description tied to tts.While I will readily admit that these tools can be useful, including them by default might induce a false sense in both developers and users that those accessibility needs have been taken care of, while in reality entrusting them to LLMs leads to countless errors that can mislead users in ways that they may not be able to discern. I often turn on subtitles when watching videos in English where the audio isn’t great or the accents are tougher for me to interpret, and the number of times I see sentences being misattributed, proper names being turned into other words or just plain misheard words changing the meaning of utterances—even in captions made by humans—is seriously worrying. I have the ability to hear what’s being said and notice the errors, but not everyone does.
(…)
Because they are accessibility features.
If someone chooses to install a similar tool, they probably understand its limitations, and quality control is on them. But if it’s provided by the OS, how clear does the OS make it that results are unreliable, and what ways does the user have to mitigate it? And does it come at the cost of implementing actual, reliable accessibility content descriptions?
How to setup OpenMW for modern Morrowind on Linux / SteamOS and Steam Deck
How to install Hollow Knight: Silksong mods on Linux, SteamOS and Steam Deck