Ubuntu’s AI Push Has Users Demanding a Kill Switch

Canonical announced this week that it’s bringing AI features to Ubuntu, and the reaction from the community has been about what you’d expect. People are asking for “a version of Ubuntu that does not include these features,” and some are already saying they’ll stick with older releases or switch distros entirely.

It’s not hard to see why. The replies under Canonical’s announcement are full of comparisons to Microsoft shoving Copilot into Windows 11, and nobody wants that kind of experience on their Linux box. The requests are straightforward: give us a kill switch, or at least a way to disable whatever AI stuff you’re planning to bolt on.

Canonical’s VP of engineering, Jon Seager, responded on Tuesday. The short version: no global AI kill switch is coming. Seager said users will be able to manage these features on a per-application basis, but that’s not the same as a single toggle that nukes the whole thing. If you’ve ever tried to hunt down every AI-related setting in Windows 11, you know why people are skeptical.

I get why Canonical wants to do this. AI is the shiny object everyone’s chasing, and Ubuntu needs to stay relevant. But the way they’re rolling this out feels like they learned nothing from Microsoft’s mistakes. The demand for a kill switch isn’t just paranoia—it’s a reasonable request from people who chose Linux specifically to avoid this kind of vendor-driven bloat.

The per-application approach might work for power users who know their way around settings, but it’s a pain for everyone else. And let’s be honest: if you’re running Ubuntu, you probably don’t want to spend your weekend figuring out which packages to remove just to get back to a clean system.

Some users are already talking about forking older Ubuntu releases or jumping to Debian, Fedora, or Arch. Canonical should be paying attention to that. Linux users have options, and they’re not shy about exercising them.

Seager’s response didn’t address the core concern: trust. People don’t want AI features that might phone home, consume resources, or change behavior without clear consent. A kill switch isn’t about being anti-AI—it’s about having control over your own machine.

Canonical still has time to course-correct. But if they keep treating this like a PR problem instead of a legitimate user concern, they’re going to lose a chunk of their base.