Pierre
Firefox for Web DevelopersAI Controls (formerly 'kill switch') are landing in today's Firefox Nightly, and will land with Firefox 148 later this month.
For the full details, see the Firefox blog https://blog.mozilla.org/en/firefox/ai-controls/
Fabio Neves 🇨🇦🇧🇷@firefoxwebdevs Great. Make it the default. Listen to the survey.
wojtek@fabio @firefoxwebdevs survey only shows a tiny fraction of the total userbase. why it should influencethe default?
aburka 🫣@aburka @fabio @firefoxwebdevs to gauge the interest of particular population and "feelings"?
this profile has less than 5k followers, participation in the survey was probably around that.
Thinking that this would impact the setup for millions of users is kinda… "cute" xD
Jakub Klawiter 
@wojtek
Because optIn is better than optOut.
@wariat @fabio @firefoxwebdevs
Yes! I know! Firefox, with each update should present a new-tab-page forcing users selecting all new additions. You know - some may not want new JS api! Or CSS adding animations! /s
@wojtek
Nope, it's not the same.
@wariat @fabio @firefoxwebdevs sure! ;)
but let's step back - why do you want it to be "opt-out"?
Eleanor Saitta@wojtek
@wariat @fabio @firefoxwebdevs Because it exposes users to a wide variety of technical and cognitive hazards, has a significant labor rights and ecological impact, is changing the structure of public discourse in ways that are broadly incompatible with democracy, and represents a fundamental change in what the piece of software the users are installing does, ontologically. Why do you think it's justified forcing that change on anyone who doesn't notice it's there or understand why they might not want it?
@wariat @fabio @firefoxwebdevs Because it exposes users to a wide variety of technical and cognitive hazards, has a significant labor rights and ecological impact, is changing the structure of public discourse in ways that are broadly incompatible with democracy, and represents a fundamental change in what the piece of software the users are installing does, ontologically. Why do you think it's justified forcing that change on anyone who doesn't notice it's there or understand why they might not want it?
@dymaxion @wariat @fabio @firefoxwebdevs considering all that - should we just ban the browser altogether?
Let's break it down:
- cognitive and technical hazzards: browser allow opening any "social"-network -- check
- public discourse -- as above
- having AI as a feature doesn't change what the software does as noone from mozilla puts a gun to user head to click on "summarise the page" button -- this is still on the user
:D
EDIT: also - should we also force Mozilla/Firefox to blacklist all "AI" websites if we are at it? Because if we don't then all those points that you mentioned can still be valid as user can just open the webpage with selected AI service… ;)
@wojtek
Right, you're clearly not interested in a conversation, so I'll let you go troll in peace
@wariat @fabio @firefoxwebdevs
Right, you're clearly not interested in a conversation, so I'll let you go troll in peace
@wariat @fabio @firefoxwebdevs
cwicseolfor@dymaxion This did not get the recognition it deserves for breadth and lucidity despite high concision, so *thank you* for putting it out there, even if the prompting party was about as receptive as a shoe-pebble.
@wojtek Interesting that you have a Vivaldi.net account, a browser that famously doesn’t include any AI features. Your question is a bit strange in this context, or maybe it’s just a bad attempt at being a contrarian.
Jake Archibald@fabio @wojtek it does include translation, which people consider to be AI https://mastodon.social/@firefoxwebdevs/115849251057488746. It even sends the text off-device to be translated, whereas Firefox does it on-device for privacy.
@jaffathecake @fabio true that… and "2,977 people" voted…
@fabio oh! what a great detective work! maybe you should dig some more to find more discrepancies? /s
I have account on vivaldi.net because I was an avid fan Opera (before they switched to chrome and then sell off to china) and when original crew started Vivaldi I got excited but it turned out it's not that great but I still got the account. I didn't use fedi all that much (tbh I started using it more only a couple of months back) but I already had an account on vivaldi so I stuck with that because "it just works".
Now, how on earth one could draw conclussion on personal perspective based only on an account domain is beyond me xDDDDD
I guess this can come only from "brave fighters" that see world black and white so in my case I should eradicate Vivaldi and this account from my computer because I wasn't completely "anti-AI"? xDDD
@wojtek bad attempt of being a contrarian it is.
Nik@wojtek @firefoxwebdevs What do you think about the ethical concerns?
@nikclayton @firefoxwebdevs of the AI? I would dance on the bonfire on bigtech ;)
alas -- I do find certain uses of LLMs (not necesarily what altman is doing) in general having some uses… that's all.
ArneBab@firefoxwebdevs that UI looks like someone really wanted to rub in the face of users "you are missing out on".
- form with long description text
- use "block" as naming instead of a plain "disable" (implies that this is a strong action on my part)
- big warning popup that names all the things again
- showing all the things anyway (but I just said no! Respect that!)
Dial FOMO to 100.
That feels really wrong.
@ArneBab "block" was picked deliberately, as it's removing all entry points to the feature, not just disabling the feature (e.g. disabled buttons are still there, but these are BLOCKED).
It was important to show the features individually so you can control them individually. E.g. translation is something people explicitly wanted to be able to re-enable https://mastodon.social/@firefoxwebdevs/115849251057488746
Ray McCarthy@firefoxwebdevs so i "re-enable" stuff I want after "blocking"?
re-enable calls for "disable" as the other part.
What that "block" does is "set default to disable" and "disable all" that we all know from the law-compliant cookie-banners.
Or rather: "make AI opt-in".
Suggestion that does not feel manipulative:
Button [disable all] ⇒ disables all AI options.
Added Option: [enable future AI features by default] ← gets disabled alongside the individual options when clicking [disable all].
@ArneBab it doesn't just disable all existing feature, it also auto-disables future AI features, or 'blocks' them from even appearing in the UI.
@firefoxwebdevs that’s what I mean with
[ ] enable future features by default.
and I’m writing my impression as feedback and this suggestion, because my reaction to the new block option was visceral.
Just seeing the workflow of enabling the block evoked the strong feeling that I’m being pulled by the nose into "do you really want to block? Why not enable just this one feature" (as drug dealers do) instead of my choice being respected.
And that does not match my usual feeling towards Firefox.
Dan Veditzsome people really like those other features though. Translation is one of our most requested features ever, and we couldn't implement prior to local AI. We weren't going to ship your web pages to some web service so they could see what you browse. But to other folks "AI is AI" and they want it off.
Feedback in Nightly is how we figure out if there are too many options or not enough.
@dveditz yes -- so why a wording that feels so massively manipulative?
For example I don’t mind image description. I don’t even mind my mastodon ALT text being mined to describe images (such a feature actually helps blind people).
But when I click "block", I mean "block" and not "Ask me whether I really want to disable all these cool features and then again whether I want to enable some selected features despite the block".
@ArneBab but it doesn't enable future feature by default.
You're blocking the entry points to these features, so they cannot be enabled.
@firefoxwebdevs enable future feature by default happens by default.
If that gets disabled on clicking [disable all] (or such), that’s the same choice, but without feeling as manipulative.
And "blocking the entry points … so they cannot be enabled" followed by options to enable them individually is part of what feels so wrong.
@ArneBab We're stumbling on the wording here. 'Enabled' to many people means the feature is active. That's why we went with 'available' for features that aren't actively running, but their entry points are available.
@firefoxwebdevs That sounds like you’re very defensive in all the hate you got the past months and made it worse in trying to give people the impression that they can make a strong statement by hitting "block".
But that’s what causes the reaction for me: if I make a strong statement, despite warning, then I mean it.
Asking "but how about this specific one" feels just like "but can a little bit really hurt?"
1/2
@firefoxwebdevs
For wording: [remove all] would match the intention you’re describing (the antonym of provide¹). And added
"default for future features: (available / unavailable)"
But not to be misunderstood:
I like it that you’re giving the option to disable all of them together. It’s just the way of implementation that feels so wrong.
It feels like FF is making a fuss about me stating my preference.
¹ https://www.powerthesaurus.org/provide/antonyms
2/2
@firefoxwebdevs And I’ve really taken up enough of your time with this.
All I wanted to do was to give you user feedback on the UI. I did that now.
I didn’t want to spread aggression (though I got too emotional) or speak against the block option in general.
I know that UI is hard and that making UI choices that touch a heated societal debate is doubly so.
To sum it up: I’m sorry for stress I caused.
/cc @dveditz
Noa@ArneBab @firefoxwebdevs @dveditz agreed with everything save for your apology. From the conversation, you were making your point rather clear & the response you got was evasiveness & gaslighting, so of course you'd get heated.
@vex I’ve been on the side of the maintainer who misjudged something and has gotten heated responses, and from that experience I think that I got too heated.
It’s really hard not to get defensive when there is a lot of heat already.
And FF devs are regularly taking the heat, even though most just try to do good work.
When you know someone receives a lot of aggression already, you shouldn’t add more.
I realized too late that we were effectively ganging up on them.
@firefoxwebdevs @dveditz
@ArneBab @firefoxwebdevs @dveditz as a former developer myself (10 years), I understand where you're coming from. Were this a normal unpopular feature, I'd agree.
But "AI" is designed to break consent from the ground up. It cannot function without theft, & there's no ambiguity around the harm of such an integration, both now & in the future.
The logical conclusion of inviting a digital bandit to the entrance of the Internet is the control of information by those who own the bandit. It has no place in a supposedly privacy-focused browser. & It's right & correct & moral to yell & gang up on anyone trying to invite the bandit to permanently gatekeep a popular entrance to the internet. Especially when the person inviting isn't naive about the consequences & is gaslighting about the what why & how.
@vex AI built on free culture content and staying true to its licenses (including attribution) is not theft.
Wikipedia is explicitly licensed to allow derivative content -- if it’s under cc by-sa, too. The same goes for almost everything I create outside my job.
Project Gutenberg provides many books that are in the public domain. Training a model on these is unproblematic.
Mozilla voice gathered voice data provided by volunteers with full consent.
@vex AI is a tool.
Most AI is trained on all the internet without consent.
But if people restrict the training set to data they have freely given consent for, then it’s morally and legally sound.
@ArneBab @firefoxwebdevs @dveditz a wolf doesn't hide amongst sheep with its claws & fangs out.
LLMs have legitimate use cases, yes. AI is an intentionally misleading label over LLMs & tend to include the less legitimate (to put it mildly) use cases. Which telegraphs intent.
Google funds something like 90% of Mozilla. There's 0 chance they're not pushing Mozilla in this direction.
@vex Mozilla has been filling up a war chest for the past years.
They can survive some years without Google.
And you say AI when you want to be understood by non-devs.
That’s not to say that you’re wrong. Just that I don’t consider your conclusion the only plausible one.
@firefoxwebdevs @dveditz
@ArneBab @firefoxwebdevs @dveditz of course it's not the only plausible outcome. "Claws retracted". But can legitimate use cases ever claw back the scale of investment in AI & turn a profit? No.
Will they stop taking Google's money? Unlikely.
Will they continue to force in the features Google says they must to continue being funded? Likely.
Will the company that dropped "Don't be evil" from their values not do evil? Unlikely.
@ArneBab @firefoxwebdevs @dveditz please don't neglect to take into account the context of these companies existing under fascism, since that context is essential. Control over information to the degree that browser integrations of AI can manage isn't something that fascism won't take advantage of. & We can already see the effects this has in media to know where this is headed.
Google is not pushing Mozilla to use AI features. To the extent people get their answers from AI rather than doing a Google search that is _costing_ us money.
@dveditz @ArneBab @firefoxwebdevs here is why I think they are:
1. Silicon valley/"big tech" pushed for fascism in politics.
2. If AI is placed at every major access point to the internet, it'll create an unprecedented amount of control over information, monopolized by its owners.
3. If AI is placed around point-of-sale access, it'll create an unprecedented amount of control over who can buy what & at what price. Even without, but especially if digital currencies become the standard.
4. The combination of 2 & 3 will allow for unprecedented population control, literally allowing those at the wheel to deny essential goods & services to the noncompliant, while rewarding compliance.
5. Fascism's wet dream is 4. Fully automated domination.
Even though you work at Mozilla, I don't expect leadership to be forthright about such a thing. I mistrust AI itself so much that I presume anyone defending it to be either a liar, or naive about how colonialism functions (& its connection to AI).
funbaker; Je suis Antifa@ArneBab @vex @firefoxwebdevs @dveditz
Dont want "AI" in my Browser is the first and only thing there is to say about this.
We all know the implications of this technology anyway, right?
@funbaker As long as the training source is used with informed consent (without coercion), I don’t mind image classification or automatic transcription.
ALT Text for images can be pretty important for blind people and replaces proprietary systems that do the same, but remotely.
For translation I’m unsure: biases AI copy from the target language are risky:
https://www.draketo.de/software/ai-translation-evaluated#completely-changed
Using AI for tab-group suggestions feels like taking a sledgehammer to crack a nut.
@vex @firefoxwebdevs @dveditz
@ArneBab people indicated pretty strongly that they wanted a way to block AI, but re-enable particular features https://mastodon.social/@firefoxwebdevs/115849251057488746.
It used to be referred to as a 'kill switch', which is much stronger wording, and a lot of folks wanted that wording (see the replies https://mastodon.social/@firefoxwebdevs/115740500373677782). But I think "block" is a reasonable middle ground. I understand that you don't.
Trey@firefoxwebdevs Great! Now make AI opt-in, not opt-out!
Michele Adduci@firefoxwebdevs please, set the block as active by default and only set AI features as "opt-in"
taylor 
@madduci @firefoxwebdevs afaik the default is already opt in in a way - for example, the link summary feature asks you if you want to enable summarisation before actually doing anything.
the point of the block is a browser wide, 'don't even ask' toggle
CubeOfCheese@firefoxwebdevs
Why not ship a version with ZERO AI?
Why not ship a version with ZERO AI?