Share your feedback on the AI services experiment in Nightly

[asafko]

Hi folks, 

In the next few days, we will start the Nightly experiment which provides easy access to AI services from the sidebar. This functionality is entirely optional, and it’s there to see if it’s a helpful addition to Firefox. It is not built into any core functionality and needs to be turned on by you to see it. 

If you want to try the experiment, activate it via Nightly Settings > Firefox Labs (please see full instructions here). 

We’d love to hear your feedback once you try out the feature, and we’re open to all your ideas and thoughts, whether it’s small tweaks to the current experience or big, creative suggestions that could boost your productivity and make accessing your favorite tools and services in Firefox even easier.

Thanks so much for helping us improve Firefox!

[Illian]

No, thank you.  Besides the issues with copyright and trademark from how they are resourced, their accuracy is negligible if not non-existent.

[Did]

Yay. I just love having glorified spambots and misinformation machines shoved down my throat in literally every aspect of modern life. Get the **bleep** out of here with this tech bro dog**bleep**. What next, a firefox brand crypto scam? NFTs?

[freezi-drink]

Please don't keep this. Firefox was one of the only places I trusted for reliable searching, and the addition of generative AI will make me stop using it.

[p4riah]

the current implementation of 'AI' in tech is a total scam. it steals. it lies. it makes things up, but it also violates everyones privacy routinely at all levels. it is harmful in the extreme.

many of us use firefox to try and claw back some semblance of privacy in our digital life. integrating 'AI' tools into the platform, no matter how 'optional', is a complete betrayal of all of your users, the platform as a whole, and everything it stands for. the fact that there are people who work at firefox who would even contemplate such a blunder is itself a betrayal. that level of incompetence should result in firings at all levels.

the disengenuous and frankly lying responses being given by mozilla employees in this thread are incredibly disturbing. you cannot enhance privacy by integrating privacy destroying tech into the platform then 'offering alternatives'. there are no 'ai' models currently in existence that even come close to being compatible with mozillas mandate. this is clearly a top-down change being forced, and the gaslighting coming from mozilla staff in here is trying to manufacture consent for a feature no one wants.

the replies basically amount to 'everyone is using ai, we must expose our entire userbase to its privacy invasiveness in order to offer them alternatives that are better' while those alternatives dont even exist. its gaslighting corporate nothing-speak that translates to 'we are adding this no matter what for shady reasons, come up with an excuse that sounds plausible and make us sound noble and the users sound reactionary'. its completely transparent and frankly, insulting in the extreme. every single person in this thread is being openly insulted by mozilla.

leave the 'AI' garbage external. let idiots and grifters install it as extensions. the rest of us want nothing to do with this poison. you add it to the software and it will rot it from within. no one will ever trust it again. i will never install a version of firefox that contains this nonsense, and i will go out of my way to tell everyone and anyone to avoid FF like the plague.

[WobblyPython]

No! God! Please! No!

[erad]

I have been actively rejecting any service that uses AI. I won't hesitate to drop firefox too.

[kaija]

This type of AI makes me sad and I think people who use it are stupid. If you make me look at another thing that pushes AI I will cry. (and look into new browsers.)

[MelMellon]

BACK OUT OF THIS NOW. A huge part of Firefox's appeal in the market is that its the morally correct choice - not for-profit, no user data hoarding, no en**bleep**tification. Implementing AI technologies - built on stealing the work of others online - is a surefire way to burn bridges with a vast majority of your userbase.

[aperkinsva7arp]

No.

Firefox is a perfectly serviceable web browser that I can use to find AI services if and when I want them.  There's no need for any kind of quick access - users can already bookmark sites and sort their bookmarks.

Also, AI trash SEO articles have already made web search practically unusable - I don't need a direct hotline to hallucinated "facts" in my web browser itself, the search engines have that part covered!

[motioti]

I do not want added AI functionality in my browser. It has already infected most websites I don’t want it

[cubiclegrrl]

Oh, HELLZ no.  Stop burning down the planet to shove unwanted bandwagon-jumping features down people's throats.  Cut it right TF out now.

[rkcurtis2597]

I think you should reconsider this idea, and would in fact caution against introducing any sort of AI programs. Generative “AI” services may be popular right now, but the downsides far outweigh the benefits. First, the amount of water and energy required for them is simply unsustainable. The use of generative “AI” has been found to be worsening climate change. This alone leads many environmentally conscious individuals, myself included, to be strongly against it. This is in addition to the fact the “AI” in its current form is not at all intelligent (hence why I’m putting quotes around it). It hallucinates information that is often untrue simply based on sequences of words it thinks should go together, with no way of fact-checking it. This can endanger users who receive this misinformation, and is a serious issue that shows no sign of improving anytime soon. Finally, Firefox has always prided itself on being privacy-focused. Many users chose it for that reason. However, “AI” programs are known to invade the privacy of internet users, “training” themselves by scraping their writing, art, and personal information without permission. Allowing these programs to be part of Firefox is counter to the goals Mozilla has stated. For these reasons, I do not believe that environmentally wasteful, misinformation-generating, privacy-invading generative “AI” technology should be part of Firefox. 

[TravestyHat]

God, no. Can you read the room? AI is a wasteful blight that makes search results worse.

[Haleigh]

No one asked for this. No one wants this. Do not add it. Simple as that.

[ETL]

NO AI. Mozilla, there are so many better features you could be developing besides jumping on the Misinformation Machine hype train.

To provide a bit nore substantive feedback: Have you noticed how every major company that introduced an AI feature had to almost immediately remove it, or heavily revamp and restrict it? GENERATIVE AI IS NOT RELIABLE. It cannot reliably summarize, it cannot reliably search, and it cannot reliably perform tasks. It's flashy and it's great at giving the appearance of sophistication, but it is absolutely money pit, both in terms of development, to try to make it do anything useful and stop producing "hallucinations" (giving a confident-sounding answer that is completely untrue), and in terms of environmental costs, as the server infrastructure required to process AI queries guzzles water and electricity.

Keep focusing on browser speed, security, and privacy. These are what draw people to Mozilla products, not hyped-up tech bubbles.

[pompom8992]

Please no! I left Chrome for this reason!

[Arina]

NO

[rhonion]

do not want anything to do with AI in my browser, optional or otherwise

[No3]

NO THANK YOU. DO NOT BECOME CHROME. I AM HERE BECAUSE YOU'RE SUPPOSED TO BE BETTER. FOR THE LOVE OF ALL THAT IS BASIC ACCURACY, RESPECT OF AUTONOMY AND HUMAN RIGHTS IN GENERAL:

DO NOT! IMPLEMENT! AI!

[Koryuu]

AI= bad

AI= inaccurate information

AI= bad for the environment

just no. No AI please

[Anonymous]

No, absolutely not! One of the main reasons I use Firefox is the lack of AI garbage! It’s a pointless, useless, environmentally damaging and entirely negative waste of time and effort, and if Firefox implements AI, I will immediately switch to a different browser service and will not return. Frankly, it’s disappointing that this is even being considered

[aima123]

If Mozilla actually cared what its users thought, it wouldn't be trying this at all. Most people I know who use Firefox switched to it to avoid AI and everything wrong with other browsers. If Firefox jumps on the AI bandwagon like this (ignoring all the ethical and practical issues with AI), then myself and a bunch of other users will just jump ship to find another browser.

[GrinningJackal]

Not only is the process of generating AI answers one which has vast and devastating ecological impact, but adding AI functionality beyond the "did you mean" spellchecker, helps confuse viewers and spread misinformation. If Firefox offers an AI option, I will be finding a new browser. I'm sure many others feel the same.

[Charleson_Mambo]

NO!

I don't want AI on Firefox.

I will not use Firefox with AI anywhere on it.

[goblinal]

No. It's not helpful, it makes everything worse.

[TheJobberwock]

Please don't add AI. PLEASE. It just eats electricity for no real benefit. And that's besides all the AIs built on mountains of stolen data.

[AlexSeanchai]

Adding long-form-autocucumber to Firefox won't be enough of a disimprovement to send me back to anything Chromium based, but it will certainly be enough of a disimprovement for me to switch to a Firefox fork that isn't sucking AI venture capital and therefore isn't inherently untrustworthy.

[Foxgloves42]

Please don't!

[Ravenesque]

there is exactly no reason to force AI into the only usable browser currently available. i would like to continue being able to tell my friends that i like Firefox and recommending they switch the way i did. if you add this AI crap i will actively discourage the switch i've been pushing among my friends.

[jaylett]

Just another Firefox user here to say don't spend time on this in future. And I'll do so via the medium of the screenshot you posted:

1. If you want a summary of a wikipedia page, you can use the Simple English variant of Wikipedia. Which, lo! has an entry for red panda. And it's better than the summary HuFace made - in part because the article seems better than the Wikipedia en original, but also because the summary lacks obvious features of a summary, such as what the body length that the tail matches actually is, that it's endangered, or anything concrete about its evolution that might justify a "taxonomy and evolution" subheading rather than just "random other facts".

2. If you thought that summarize was a good feature, then why would you crowd so much of that output with AI crap around the edges? If I want a summary and you think I'll be happy with the quality of this output, why should I care that you generated it by doing some kind of prompt dance? Why would I want to upload a file as a next action? What would that even do? What are the six tools? If any of them is useful, why aren't they available as a toolbar of some form? I can't peer inside the mind of your product designer, but my assumption is because you decided to add a chatbot to Firefox, and that's what a chatbot looks like for some reason. Identify a user problem. Solve the user problem. Don't trip over the furniture.

3. What is that headline doing? "Red Panda - A Small Mammal with a Big Personality". The LLM has just made that up presumably because "a small X with a big personality" is a moderately common snowclone.

The screenshot doesn't sell this feature, it should bury it. It looks like a quick proof of concept that someone threw together because their boss had read that McKinsey thinks that Krug comes out of Sam Altman's nipples. Mozilla is supposed to be better than this.

[MelMellon]

Incredibly well done summary of all that's wrong with this

[samarkandar]

Not only should Mozilla not implement AI, you should fire whatever lemming-brained empty suit suggested doing so.

[ajet]

Absolutely not. I personally think theres a time and place for ai, but I'm sick of scrolling past the ai results on google everytime I look something up because the results are so often wrong or even dangerous. I'm sick of instagram having meta ai in the search bar...and for what!? What does the ai do for me that simply searching for keywords don't - on instagram, at least? If I have to double check everything the ai tells me, its ultimately not even saving me the time of doing my own research or writing.

Is its usage worth the energy costs? How long will these tools be sustainable, both financially and environmentally? Will they have to eventually implement pay walls like so many other ai tools out there?

Have heard stories of younger people now using ai instead of search engines, and taking complete ai hallucinations as fact because they believe the ai is infallible. Is this really worth it? I turned to Firefox specifically because it does not have the same BS Chrome does.

I understand this is intended to stay optional, and an effort to stay relevant, but I fear it will become more and more integrated like it has for every platform.

AI is potentially an amazing tool for recognizing patterns in medical scans or research photos, heck, in the early days I saw it being used for a tool to help blind people identify surroundings via their phone camera. Now I feel it's become a marketing gimmick that is solving zero problems. Eventually the sparkle will fade and we'll be left with the wreckage mass misinformation and energy overconsumption caused.

[Ark]

Just do not do this. Not ever, not for any reason. Don't.

[webenjoyer]

This message is for @Mardak and anyone else at Mozilla responding to this obvious headache. I'm taking a look at your replies and I must give you some frank advice.

You're asking questions like "Are there ___ to make ___ better?" and "Would there..." People can see that you are trying to use reframing speech and you need to stop and take a step back. You're trying to reframe this by finding specific aspects of the feature to make this seem more palatable. I see this a lot from AI enthusiasts. From one software professional to another, I'm here to tell you that there isn't any one aspect of this feature that is the problem. The feature is the problem. It happens sometimes. You are focused on a second-order problem when the first-order problem is right there.

As a software professional for over 15 years, I give you the following advice.

Imagine a product that does bad behavior X. To make the product better, you remove X. Sometimes X is "leaks the user's information" or X is "a vulnerability". Now, imagine a product that does bad behavior X with an additional bad aspect of Y. We don't focus on fixing Y, we focus on X. I hope you can understand that.

Privacy is not the problem with LLMs, although it is a major problem, because privacy a fundamental problem. I'm telling you that LLMs are fundamentally problematic in the same way as poor privacy practices. LLMs are more fundamentally problematic than pretty much anything you can imagine: security, privacy, performance, reliability, you name it. It is attacking the very reason for being alive and thinking about any of these problems in the first place.

I mean, imagine that everything you've ever been interested in or made in your life was taken, analyzed, and recreated with subtle tweaks by anyone at the click of a button. Multiply that by every word or piece of art that all of humanity has made. Everything is reduced to a spreadsheet of vectors. Now, we market that into a product that people pay for. Even if it were free, don't you see how messed up that is? We should be spending efforts on education, promoting arts and individual expression, giving people reasons for going out into the world and being influenced by it in a nuanced way. And then letting them struggle to remix it in their own unique way, EVEN IF the "product" is "not good".

Machine learning can be such a gift of automation--in medicine and other methods of broad analysis, in dangerous and tedious work. These are things that we may not have the time to spend on or things that we don't want to spend our time on.

LLMs are completely misguided because they automate the parts of life that we are trying to spend our time on! Writing, art, expression: that's the good stuff! Don't take that away.

If you can't understand all of that in a deep, intuitive way, such that it seems obvious to you, you need to take a step back from computers for a few years until you realize what's really important in all of this work. I can practically feel your discomfort from here with this absolute deluge of negative feedback, and I feel for you. What might seem like the noise floor to you might actually be something more like the cosmic microwave background. Not something to ignore. I give this feedback to you as a friend to the world. Hope that makes sense. Take care.

[lovelyjubbly]

The floating AI icon that appears (sparkly emoji), feels like it appears too readily. On some pages, any kind of dragging or moving action produces it. It would be nice if it stuck to the context menus, or was a bit more discerning, for example should it really appear if a single word was highlighted? I think it does still have usefulness appearing immediately after highlighting a paragraph though, and it's also a way of promoting that this feature exists to us users.

[trer_]

Hi Mozilla, this is a terrible idea.

(I'm not going to "try out the feature", the mere idea of indulging that request abhors me)

Your web browser keeps existing because people are passionate about having an alternative to the crap that the likes of Google, Microsoft, and Amazon are throwing at us. These companies care not about their users and you cannot compete with them on their playing field. The only reason you keep existing is because people trust you to care more about them, their freedom, and their safety than about short term profits.

Please don't keep gambling that away on hype and buzzwords.

[Hjalti]

I do not want AI services in the sidebar. I do not want AI services anywhere to do with Firefox or any of Mozilla's other software.

[DamienW]

Extremely disappointed to see Mozilla partnering with multiple anti-privacy, anti-user, and environmentally damaging companies to integrate tools and systems which reinforce a paradigm of manufactured, coerced consent to people's choices over whether and how their data is used. The values embedded in the operations of LLMs and other "generative AI," as presently construed and created, predominantly sit in direct opposition to everything Mozilla has been understood to work towards and hold as Good.

If this were a project about Mozilla working to building a better, more ethical, consent-focused, environmentally sustainable LLM or "GenAI" system, then I would support it without reservation; more than that, I would applaud it and point to it as a meaningful change in the status quo and a good path forward in the world of "AI." As it stands, this just reifies the worst tendencies of "AI" companies, casting them as "inevitable" or otherwise without alternative; Mozilla of all places should have the vision to resist these PR lines about "AI," and at least try to imagine a different way forward.

[threeley]

Absolutely not.