Firefox uses on-device downloaded-on-demand ML models for privacy-preserving translation.
-
@joepie91 @firefoxwebdevs Mozilla's tortured definition of opt-in seems to predict that Mozilla will invent features to nag you into enabling AI, as they have already done with Link Previews: https://www.quippd.com/writing/2026/01/06/architecting-consent-for-ai-deceptive-patterns-in-firefox-link-previews.html
-
The Firefox AI "kill switch" is not "complicated" except insofar as it's incoherent. it's not "undisclosed nuance" except insofar as it's incoherent.
the "kill switch" doesn't exist.
this is important to keep in mind. once you remember that NONE OF THIS EXISTS, you will realise that every one of the dilemmas you posit is an imaginary problem that follows from incoherent postulates.
e.g. "AI kill switch purists" is not a coherent postulation because the "kill switch" does not exist.
the "kill switch" is a hypothetical proposed in this post:
https://mastodon.social/@firefoxwebdevs/115740500373677782
the "kill switch" is a proposal to satisfy the demand for an opt-in by providing an opt-out. you might think that's a failure to respect the question, and you might even begin to suspect the proposal was in bad faith.
note that Jake, in presenting the kill switch and calling it a kill switch and getting it into all the papers as a kill switch, says he's uncomfortable with the name he's publicised it as. you might think that's oddly incompetent for literally a PR (devrel) person.
the concept as presented imposes multiple false dilemmas.
the LLM stuff should *incredibly obviously* be an extension. this is the purest possible opt-in, despite jake's past attempts to muddy the meaning of "opt-in".
making it an extension is also eminently feasible. There is literally no technical reason it needs to be a browser built-in.
this suggests the reasons are not in any way technical. some person with a name, who has yet to be named, dictated that it would be a built-in. so that's what Mozilla is going with.
why Mozilla went hard AI is entirely unclear. this would have been late 2024? we have no idea who was inspired with this bad idea nor why they were so incredibly keen to force it into the browser.
nor is it clear what Mozilla will do for external LLM services when the AI bubble runs out of venture capital and pops in a year or so, most of the chatbot APIs shut down and whatever remains is 10x the cost at least. but that's a problem for 2027's bonus, not 2026's.
note how the poll provides no option for "no LLM functions built-in to Firefox", in a pathetically transparent attempt to synthesize consent. jake wants to use this poll as evidence of what the user base wants, deliberately leaving out the option he knows directly a lot of them want.
and in conclusion:
1. solve the "kill switch" naming problem by branding it the "brutal and bloody robot murder switch with an option on the executives responsible".
2. make all this shit an extension like they should have a year ago.
3. and your little translator too.@davidgerard @mdavis @firefoxwebdevs where did I say I'm uncomfortable with the name "kill switch"?
-
@firefoxwebdevs I really love the local on-device translation, "AI" or not.
I think this question follows a fundamental misunderstanding of the AI toggle. I want I do not want to ship off my browser data to any AI company (including Mozzila), and that would be the toggle I would look for.
If Firefox/Mozilla came out with a on-device local-only LLM I would personally be more receptive. The main issue for a browser is that it should be a browser, and also not ship all my data off for harvesting by AI slop companies.
@soupglasses I agree with your take here, but many people in the replies have a more fundamental dislike of 'AI'.
-
Firefox uses on-device downloaded-on-demand ML models for privacy-preserving translation.
They're not LLMs. They're trained on open data.
Should translation be disabled if the AI 'kill switch' is active?
@firefoxwebdevs I voted "no" because I'd agree - this shouldn't be considered the toxic "AI".
-
@chillicampari @firefoxwebdevs @joepie91 I’m kindof amazed that Mozilla can’t distinguish which changes led to the backlash. I think that’s why this whole thing feels more like putting on a show than like a genuine attempt at reform.
The timing alone makes it clear that the builtin translation was not the issue. Sure, moving it to a plugin would be an improvement, and requiring user action to enable it would be smaller improvement, but that was the case before.
️@ShadSterling @chillicampari @firefoxwebdevs @joepie91 translation is already opt-in. You're prompted about it, and the model is only downloaded if you say you want it.
-
@firefoxwebdevs doing a great job at regaining users' trust there, I see
In other news, you've done such a great job at regaining my trust that I've switched browsers to anything but Firefox. Well done, Mozilla.
@mxjaygrant what was it about this post that made you switch?
-
@firefoxwebdevs Also as a side note: The org I'm working on has banned genAI tools for projects above a certain level of confidentiality. Guess what? Firefox is banned as well and probably stays banned regardless of any kill switch.
@sebastian which feature resulted in the ban? Given that you can access eg chatgpt in any browser, shouldn't your company ban all browsers?
-
@firefoxwebdevs The frame of this question is risible.
I am begging you to just make a web browser.
Make it the best browser for the open web. Make it a browser that empowers individuals. Make it a browser that defends users against threats.
Do not make a search engine. Do not make a translation engine. Do not make a webpage summariser. Do not make a front-end for an LLM. Do not make a client-side LLM.
Just. Make. A. Web. Browser.
Please.
-
Firefox uses on-device downloaded-on-demand ML models for privacy-preserving translation.
They're not LLMs. They're trained on open data.
Should translation be disabled if the AI 'kill switch' is active?
@firefoxwebdevs alternative perspective:
Remove all AI-LLM, AI-ML related functionality.
Then have target end-user (web developer) choose, informed by their values & preferences what functional components they’d like to “plug-in” to web-browser for ML content processing for web page-
- Language translation - enable on device locally download-on-demand ML or use your own
- Dictionaries
- …
Once these are real-world validated & functional, they can be shared via open source commons with others. -
@sebastian which feature resulted in the ban? Given that you can access eg chatgpt in any browser, shouldn't your company ban all browsers?
@jaffathecake ChatGPT (and many other web based things) are firewalled.
Also you are looking at a compliance issue from a technical viewpoint. As the implications of genAI generated content wrt. copyright and things like patent applications are still somewhat unclear in many jurisdictions, the simplest solution is to stay well clear of any tool that claims to do anything "AI".
If the contract with the customer says "no AI because it exposes us to legal risks", then the work has to be done in a clean environment where there is nothing that could be considered AI. -
@ShadSterling @chillicampari @firefoxwebdevs @joepie91 translation is already opt-in. You're prompted about it, and the model is only downloaded if you say you want it.
@ShadSterling @chillicampari @firefoxwebdevs @joepie91
You're *constantly* prompted about it on every single site you visit. Calling that opt-in stretches the definition of consent.
-
Firefox uses on-device downloaded-on-demand ML models for privacy-preserving translation.
They're not LLMs. They're trained on open data.
Should translation be disabled if the AI 'kill switch' is active?
@firefoxwebdevs I would rather like for auxiliary features to be added via the extensions API.
-
Firefox uses on-device downloaded-on-demand ML models for privacy-preserving translation.
They're not LLMs. They're trained on open data.
Should translation be disabled if the AI 'kill switch' is active?
@firefoxwebdevs There's nothing intrinsically wrong with AI. If you can do translation on device in a privacy-preserving way, there's no reason to disable it.
-
Firefox uses on-device downloaded-on-demand ML models for privacy-preserving translation.
They're not LLMs. They're trained on open data.
Should translation be disabled if the AI 'kill switch' is active?
-
@firefoxwebdevs I'm not a fan of this all or nothing approach to functionality. What happened to the paradigm of things being scope and/or time-limited, eg just for this page, for x minutes?
@scribe it isn't all or nothing (the currently winning option is granular control). This is the first time I've heard a request for AI to be disabled in a time-limited way. Tell me more about your use-case.
-
@scribe it isn't all or nothing (the currently winning option is granular control). This is the first time I've heard a request for AI to be disabled in a time-limited way. Tell me more about your use-case.
@firefoxwebdevs I think it's less about use cases and more about general trust, as privacy often boils down to. If you're talking about an AI "kill switch", you're talking about trust in what's been defined as "AI", and trust in the browser developer as a whole.
Once definitions are murky, there's an area open for ongoing redefinition. One way to adopt a "private by default" approach is to follow what cookies do, for instance, and allow users to allow limits to the extents of permissions.
-
Firefox uses on-device downloaded-on-demand ML models for privacy-preserving translation.
They're not LLMs. They're trained on open data.
Should translation be disabled if the AI 'kill switch' is active?
@firefoxwebdevs there's a huge difference between calling out to an external llm service and using the translation-specific on-device models.
(If, hypothetically, llm's like chatgpt were not a thing, would people have such visceral reactions against the translatelocally models?)
-
@firefoxwebdevs there's a huge difference between calling out to an external llm service and using the translation-specific on-device models.
(If, hypothetically, llm's like chatgpt were not a thing, would people have such visceral reactions against the translatelocally models?)
@unhammer a lot of folks in the replies & responders to the poll feel differently. I personally agree with you, but I want the kill switch to have broader appeal.
-
@davidgerard @mdavis @firefoxwebdevs where did I say I'm uncomfortable with the name "kill switch"?
@jaffathecake @mdavis @firefoxwebdevs in the quoted post included as the reference
-
@firefoxwebdevs I think it's less about use cases and more about general trust, as privacy often boils down to. If you're talking about an AI "kill switch", you're talking about trust in what's been defined as "AI", and trust in the browser developer as a whole.
Once definitions are murky, there's an area open for ongoing redefinition. One way to adopt a "private by default" approach is to follow what cookies do, for instance, and allow users to allow limits to the extents of permissions.
@scribe I'm not sure cookies are a great source of inspiration when it comes to privacy
