Following yesterday's polarizing news of Canonical to begin shipping AI features in Ubuntu Linux over the course of the next year, Jon Seager as the VP of Engineering at Canonical has now provided some clarifications around their AI plans.

A short time ago Seager posted an update to his Ubuntu Discourse announcement around rolling out of AI features in Ubuntu over the next year, beginning with Ubuntu 26.10. He still has yet to elaborate on possible AI-driven features for Ubuntu Linux but at least clarified that it will initially be opt-in and in effect the "AI kill switch" for Ubuntu will be a matter of just removing some Snaps.

Seager explained in this morning's update while there won't be a "global kill switch" for AI features on Ubuntu, their plan is to deliver all the AI features via Snap packages. So removing AI features on Ubuntu will come down to removing Snaps.

The AI-backed features will initially be as a "preview" and "strictly opt-in basis" for Ubuntu 26.10. For Ubuntu 27.04 and beyond, the plan is to have an initial setup wizard around prompt users on the AI native features they may want or not. Part of this is being done since due to the large size of LLMs, they won't be baked into the Ubuntu install ISOs but have to be downloaded after the fact and thus can be punted off to an initial configuration wizard.

For addressing concerns around any cloud-based AI usage, Seager noted that those aren't part of the plans. The default configuration "will always be" to use local inference against local models.

He also did elaborate as some possible AI features like camera focus and text to speech:

"The plan here is not to force AI into every Desktop indiscriminately, but rather to enhance certain features where it makes sense (text to speech, camera focus, etc.) and make it easy for those who do want to consume AI features to do so in a way that they can trust, and that is well-integrated."

Lastly, the VP of Engineering did clarify that they will be shipping code in Ubuntu that is co-authored by AI.

"Regarding whether or not we should ship code that was co-authored by an AI - in reality we will be doing this. Even foundational projects in the ecosystem like the Kernel itself now have policies around how to govern this, and will accept tasteful, correct contributions that have been authored with AI."

Again, all of the official commentary in full can be found via Ubuntu Discourse.