33

I'm more asking this on behalf* of the many people I see trying to express protest to the OpenAI partnership in ways that get themselves into trouble:

  • Deleting their own posts- individual or en masse
  • Defacing/vandalizing their own posts- subtly or non-subtly

The result of doing those things is those users just get themselves into trouble because it's vandalism. See /help/what-to-do-instead-of-deleting-question. Nor is it effective because the system is designed with rate limits and automatic-flags to prevent or slow down such deletions from people "rage-quitting", and there are user-operated systems to detect such vandalism. I won't make an ethical argument, but I will comment that I think it's sad that people resort to destructive actions on this resource (especially when they themselves care about it).

Lots of people are suggesting these ineffective and non-constructive methods of protest both in the OpenAI partnership announcement post, and off-site. In no coincidence, a lot of people are trying these things, and facing the consequences- especially on Stack Overflow right now (which has not gone unnoticed by the "outside world", and which I don't think is good PR for what is supposed to be protection of this community resource).

People can also delete their accounts, but that's pretty nuclear. You can't go back after that, and it seems like a lost opportunity to leave a message to future network users about why you did it. It doesn't cause your posts to get deleted- just semi-dissociated.

What can these people do to express their protest in a way that won't get them into trouble, and ideally is effective? (communicates clearly what they are unhappy about and has reach).


*I haven't found reason to protest at this moment because the announcement has no substance of detail about what the partnership actually means (what license is on the subscriber content given to the partners, what changes will be made to the SE platform, etc.), though I respect other peoples' choice of what they want to protest about.

CC BY-SA 4.0
5
  • 2
    Are rational, non-aggressive and fact-based posts that express users' feelings, published on the company's META sites, a valid and constructive way of protesting? Or are they just the focus of unnecessary discussions? Not that I'm thinking of publishing something like this. yesterday
  • 12
    @AugustoVasques That's literally the point of meta IMO - its part of the cycle of governance that SE should have for a healthy company/community relationship. yesterday
  • Just thinking aloud 𝚆𝚑𝚊𝚝 𝚊𝚋𝚘𝚞𝚝 𝚞𝚜𝚒𝚗𝚐 𝚞𝚗𝚒𝚌𝚘𝚍𝚎 𝚌𝚑𝚊𝚛𝚊𝚌𝚝𝚎𝚛𝚜 𝚒𝚗𝚜𝚝𝚎𝚊𝚍 𝚛𝚎𝚐𝚞𝚕𝚊𝚛 𝚌𝚑𝚊𝚛𝚊𝚌𝚝𝚎𝚛𝚜? Several alphanumeric Unicode characters are readable by people as regular / ASCII characters. I have yet to learn how screen readers will work with posts using this kind of character, but they have been used for a long time on social networks. One of the caveats is that posts using them might be automatically flagged as suspicious, so there is a risk regarding how the posts will be handled by reviewers / mods / CMs.
    – Rubén
    yesterday
  • @Rubén please put answers in an answer post- not in comments. yesterday
  • It's just a thought, not an answer, at least it's not ready to be posted as one.
    – Rubén
    yesterday

7 Answers 7

17

Anecdotally a lot of the folks involved seem to be folks with older accounts and somewhat underused. It also feels to a certain extent to be emotive. A lot of these folks haven't been active for years and seem baffled at moderation actions that have been done this way for years.

I've often found that people feel anger at the wrong people - sometimes lashing out at moderators or community members trying to maintain what are essentially long standing rules against self destruction of their posts, or even getting a badge, cause apparently they don't understand this is how we've always done things.

Rather than trying to burn down this place and the posts/content here, a productive way to redirect that anger might be to go back to the 'human' network. Work with sites that don't tie up with genAI companies. Work on your blog (and make sure you have the no AI robots.txt entries).

I'm personally convinced genAI is a grift, and that bubble will burst eventually. In the meanwhile I'd like folks to consider their actions and use that anger in a more productive way. Ragebuild, don't rageburn.

Deleting posts doesn't help cause they're probably already scraped/scrapable. Poisoning your code doesn't help cause outside the most trivial examples, genAI does a bad job at it anyway, hallucinating libraries and such. It might feel like you're doing something but I'm doubtful.

CC BY-SA 4.0
12
  • 2
    What is genAI? Generative AI? In any case, AI is no "grift", it's like someone saying in 1990 that the Internet was overrated. AI is here to stay, for better or for worse. yesterday
  • 15
    The idea of it ? Not really. Companies like openAI claiming it'll solve everything and replace a whole range of human artistic, creative and skilled labour? That's a grift. AI is here to stay when its boring and invisible. Also, I'm stating it as an opinion, and that's potentially wrong. I've a very mixed track record on these things, from predicting Gamin should be a wearables company in a term paper(I failed that paper, but turns out I was right) to an aweful lot of tech I think will be cool but end up niche :D. yesterday
  • @Mari-Lou I have no doubt certain industries will be irrevocably changed by the use of AI. From Help Desks/Support systems to Data entry, Creative spaces (both writing and imagery), as well as integrating into search and digital assistants. But not at the level or scope certain companies in the space are claiming.
    – Robotnik
    yesterday
  • Artificial intelligence isn't something new, it's been around since 1966, the only difference is that today the chatbot is accessible to anyone on the planet who has a smartphone or an internet connection. ChatGPT and its rivals are nothing new, in fact they are the third generation of AI bots. There was already the first generation chatbot ELIZA and the second generation chatbot ALICE. yesterday
  • 1
    ChatGPT is only the forerunner but just like Ask Jeeves was the best search engine in the mid 1990s until Google arrived. Do young people today even know about MySpace? No, but they certainly know and use Instagram and Tik Tok. Suggesting that AI is a grift is wishful thinking, its scope will expand, it will permeate education, the Internet, and work. yesterday
  • 1
    @Mari-LouAСлаваУкраїні AI is a very broad term. The current fad is about large language models and they, by definition, are limited to their training data and doomed to repeat common errors in said data. That, I also think is a fad and likely short lived. General AI is a whole different thing.
    – terdon
    yesterday
  • 1
    @Mari-LouAСлаваУкраїні, Altavista was better than Ask Jeeves. yesterday
  • 1
    @terdon I was obviously referring to generative AI, i.e ChatGpt3.5 and 4.0, Dall-E 2, Gemini and its ilk as I presume was Journeyman. The potential is huge, it won't all be bad but a lot will (says the cynic in me). yesterday
  • Yes, @Mari-LouAСлаваУкраїні and that (generative based on LLMs) is what I also think will be relatively short lived because of how limited it is. My bet is that it will soon(ish) be replaced by more powerful approaches.
    – terdon
    yesterday
  • 1
    @terdon Like Google replaced Ask Jeeves and Altavista. My point being that LLM or GenAI etc. are not going away. They will be replaced, but still this creativity, problem solving programme will continue to evolve. Come back in 5 year's time and tell me Journeyman was right. yesterday
  • 1
    Anything with a bunch of hype attracting investment money has a bunch of grifters attached to it. The hype will die down, a bunch of people will lose a bunch of money, but the technology will remain. We’re still in the phase where we figure out exactly what it’s good for. It makes me sad, but I don’t care enough about SE any longer to want to spend my energy protesting their intransigent short-sightedness.
    – ColleenV
    7 hours ago
  • 1
    More or less. Good, useful technology is boring - but I'm cynical about the long term as well as how the current genAI industry is literally mirroring crypto - throwing massive amounts of computing power at a problem that still feel almost like a chatbot or glorified autocomplete. 7 hours ago
12

How can a user unhappy with the LLM partnerships protest constructively (and ideally- effectively)?

A constructive protest implies that there's at least one outcome that can be achieved which would appease the protester, and allow them to stop protesting. I think it's worth pointing out - some of the users who are defacing their own posts are not amenable to any sort of outcome. They've been wronged, and are fine burning the bridge behind them. For those users, the best we can offer is a "thanks for your time and good luck with your future endeavours".

Organise

For those frustrated with the current situation but open to real change, then the best course of action is to organise and collaborate with other people, on or off site. Figure out what goal you are trying to achieve, what's the ideal outcome and what are the alternative outcomes, and compromises the group is willing/not willing to accept.

Once you have that - a combined vision and game plan, the next step is to choose your actions: username/profile changes, open letters, meta discussion threads, strikes, or even simply leaving the site en-masse, are all options that - with a large enough base, can cause change to happen. This is ultimately how the mod and community strikes worked, and at a basic level, is how unions work.

CC BY-SA 4.0
8
  • 1
    On the current issue of OpenAI - I would accept an "opt-in/out" of data harvesting for AI purposes at the user level - this would mean any content produced by me and shared on SE would be unavailable to Open AI. I would prefer this to be an opt-in system, but that is too much to ask in this capitalistic hellscape of 2024 I fear (:
    – Robotnik
    yesterday
  • Part of the problem is the lack of trust. Even if SEI promised us opt-in... there's no reason to trust they'd keep that promise. yesterday
  • @S.L.Barthisoncodidact.com I'll trust it once it hits the legal/TOS/Privacy policy. Before then it's just words
    – Robotnik
    yesterday
  • "For those frustrated with the current situation but open to real change, then the best course of action is to organise " - been there, done that, didn't work
    – Zoe
    yesterday
  • @Zoe It achieved many of its aims. We now know that those aims didn't serve our goals as well as we would've liked. Different aims might have a different outcome.
    – wizzwizz4
    yesterday
  • 1
    You left out "hire an international intellectual property lawyer to figure out how to get users out of the license agreement, or tie up the material in question for as long as possible in litigation". yesterday
  • 2
    @wizzwizz4 It worked to the point only because almost all moderators on SO and maintainers of the spam fighting bots and the most active curators supported it. In this situation, I don't see that happening. We already had the "selling data to AI" discussion last year, the current wave of protesters are mostly people who have been long inactive on sites and have been suddenly stirred up by some half truths picked up on Mastodon or Twitter. Active users have different goals. Not that we are thrilled about AI, but our expectations are grounded in reality. 12 hours ago
  • @Robotnik Such an opt-out is fundamentally at odds with the concept of open content and, definitely, against the CC BY-SA license. The CC BY-SA license provides no way for the copyright holder to restrict access to only specific purposes. It would be possible to have SO, the company, provide an opt-out from the data they sell, but that's not reasonable, because as soon as that happens the LLM AI companies would just go back to getting the data from the data dump, SE API, or scraping, which would then push back onto each individual contributor the enforcement of the license terms.
    – Makyen
    5 hours ago
6

Contribute to an alternative Q&A site or content fork

The main weapon for any open source project unhappy with its management is to fork the codebase. There are many examples of code projects doing this.

The equivalent for a content-based project like SE or Wikipedia is to take a snapshot of the content, host it elsewhere, and then build on it.

This achieves two objectives:

  1. It moves (part of) the community away from SE.
  2. It duplicates the content that SE currently holds, making it valueless for SE to sell to AI companies.

A successful fork leaves the original management with nothing but a snapshot of content that is also hosted elsewhere and a dwindling community.

Often, the threat of a fork is enough to keep the management listening to the community. It's likely that the threat of forks is what kept Wikipedia largely on the straight-and-narrow in the early days.

Where can we go?

There isn't any true fork of Stack Exchange currently running. You can try to organize and get this together, or just focus on aim 1 for now, and take your contributions elsewhere.

Codidact is one alternative project that has experimented with importing SE content. It currently doesn't aim to become a full fork, for various reasons, but perhaps with some extra community contributions, the idea could be investigated again.

Another alternative which aims for a less profit-driven managementis topanswers.xyz. It doesn't rehost SE content, but it still achieves aim 1.

Show people where you've gone

Once you've found an alternative community, you can use a simple addition to your username to signal that you've left and where you've gone. The Monica Cellio affair shows that if enough people do this, it generates quite a bit of attention for an issue.

CC BY-SA 4.0
New contributor
Peter is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
15
  • 1
    In the broader sense that Wikipedia is an open source project. It's a large community created resource, under a copyleft license, with (originally). In any case, the ability to fork was explicitly built into SE/SO from the start for just this purpose.
    – Peter
    yesterday
  • 12
    Codidact is not a fork, it is an alternative Q&A site with its own code base. yesterday
  • @samcarter_is_at_topanswers.xyz It's more of a partial fork. It contains some of the content from StackExchange, but only in piecemeal. With some effort it could become a complete fork.
    – Peter
    yesterday
  • 4
    Note that I'm talking about forking the Q&A content, not the stackexchange codebase (which isn't open source).
    – Peter
    yesterday
  • 5
    In this case, using the word "fork" is incredible confusing. It makes one immediately think of the code base. yesterday
  • 2
    Please note that importing all the content is currently not the aim of Codidact. Codidact experimented with this for some communities, but this lead to a lot of death and uncurated content and also seems bad for SEO. yesterday
  • @samcarter_is_at_topanswers.xyz Your edit is partly missing the point. It's not just to move the community to another site, but also to host and build on a snapshot of the existing content. The aim is not just to move away from the SE management, but to make their current content valueless, since it is also hosted on the alternative site.
    – Peter
    yesterday
  • @Peter If that's your aim, I'm not sure why you mention Codidact. This might lead to wrong expectations. yesterday
  • 2
    I've rewritten the answer to (hopefully) focus the attention on the idea of a content fork, and to highlight the two separate aims.
    – Peter
    yesterday
  • How about joining the IndieWeb movement? People should start by getting a personal domain and learning about using web mentions among other stuff.
    – Rubén
    yesterday
  • 1
    "to host and build on a snapshot of the existing content" AFAIK, scrapers have done that goal far before the recent debacle... yesterday
  • @MetaAndrewT. True, but that's usually a cynical push to grab some clicks, and the result cannot compete in any way with the original. The aim of a fork would be to take the content, and a substantial part of the community and continue under different management, so that the new project is more active, and better maintained than the old.
    – Peter
    yesterday
  • 1
    In what way does contributing elsewhere affect anything? the important part is simply walking away from here.
    – Kevin B
    yesterday
  • @Peter ah, right, I forgot that the goal is not about SEO, but about the value of selling content. Indeed, someone mentioned that Codidact still hasn't blocked GPTBot for now (serious comment, not a sarcasm) yesterday
  • 1
    @Rubén: Indeed they should. But it is too convenient to use the centralised services for most people. The cost is on the order of one coffee per month (and the minor hassle of renewing, etc.). Centralised services can wipe out all your content in an instant, with very little recourse (at the very least keep a backup (of important content)). yesterday
5

You can change your username to something that expresses your dissatisfaction. Of course, it needs to be something that is appropriate according to the Code of Conduct and other site rules (Ex. no expletives). This way, every post you've written will send some message of your choosing. This has been and still is a popular method of protest to unpopular company decisions.

You can also elaborate on the reason for your dissatisfaction in your user profile's "About Me" section.

Expressing yourself (respectfully) in an answer to related announcement posts on meta is also a good option. Don't forget to vote up other existing answer posts if you agree with them.

You can protest off-site too. There was a point where SO Inc. apparently seemed to listen to people on Twitter, but I don't know how true that is now.

Here I make my personal plea that whatever message you send, hopefully you convey that the concept of sites like the Stack Exchange network sites is still useful and worth pursuing (if you already/still believe that).

CC BY-SA 4.0
4
  • 4
    The whole current mess started on Mastodon, and later was reposted on twitter. TDD makes us sad. yesterday
  • I thought you wanted effective ways of protest? Changing usernames and expressing dissatisfaction in the "About Me" section are anything but (see also: Monica). At this point, I don't think the company will care unless the chosen protest method impacts the company revenue or compromises some aspect of SE (e.g. answer quality) enought to cause significant problems.
    – l4mpi
    yesterday
  • Usernames don’t really work except for winks and nods from like-minded users. The company doesn’t care what your username says, and the LLM sure doesn’t care. If you have answers on technical sites, maybe jamming key words or phrases into method or variable names would make that message get relayed more broadly (e.g. in any code LLM “produces” based on your answer)? Subliminal, still, but potentially wider-reaching. yesterday
  • 3
5

Here are two simple ideas. The main issue with both is that they are not directly communicating to SE that you are unhappy about AI, but they both have the potential to actually affect the company.

#1: Install an adblocker, or if you already have one and whitelisted SE, remove it from the whitelist. Tell everyone you know who uses any SE sites to do the same.

This costs the company money, and thus at least has a measurable effect, even though it is small for an individual user. Simply expressing your dissatisfaction in writing (meta answers, your profile, etc.) doesn't have any effect in most cases, as was demonstrated lots of times during the controversies of the last years.

#2: Another potential option might be to submit requests that can only be handled by CMs or other company representatives and not by moderators - effectively creating extra work for the company. Note that I am not advocating spamming SE with illegitimate requests, but there are requests to which (at least some) users are legally entitled.

For example, EU citizens have the right to perform various data requests under the GDPR which must be answered in a timely manner, including requesting an export of your personal data in a machine-readable format (note the GDPR only covers your personal data and not your questions or answers). I'm not sure if some of those requests are handled by fully automatic processes, but if they require any form of work from a company employee, the company would definitely notice if they suddenly received lots of these requests, as ignoring them or taking too long is not an option for legal reasons.

And a bonus #3: Stop contributing, stop moderating, stop flagging, and stop doing any free work for this company. For example, don't help SE cleaning up after users who deface their posts, let the company do it by themselves (see also this answer by Shog9). This is significantly more effective if you're a moderator and go on strike, but if enough regular users stop doing any curation work this will eventually have an effect as well.

CC BY-SA 4.0
0
4

I am not an active user on SE or SO, but I often participate by reading the Q&A, particularly in my fields of interest such as embedded & C, and it is something that really brings me joy the amount of knowledge I have learned so far with shared content.

After the moment that Partnership with OpenAI was announced, I noticed most of the users, the most active and high-reputation ones had stopped to answer or comment, leaving many unanswered questions.

Maybe it is a kind of protest or they just have abandoned SO for good, this is something I cannot answer. However, it's clear that current and future users are the ones losing out in this situation.

Something that I can imagine is the creation of an open-letter with the digital signature of the SE users to express their unhappiness with the current situation, explaining in detail why they are against this change and how the partneship with OpenAI might affect the SE, and sharing it on social media like Twitter "X" and LinkedIn to expose the current situation and how it will influence SO onwards.

Here I make my personal plea that whatever message you send, hopefully you convey that the concept of sites like the Stack Exchange network sites is still useful and worth pursuing (if you already/still believe that).

CC BY-SA 4.0
New contributor
Geisterfahrer is a new contributor to this site. Take care in asking for clarification, commenting, and answering. Check out our Code of Conduct.
3
  • 5
    "After the moment that Partnership with OpenAI was announced, I noticed most of the users, the most active and high-reputation ones had stopped to answer or comment, leaving many unanswered questions." Citation needed. I think there's more bemusement and befuddlement and there's a interesting trend of users who have old, barely used accounts doing this rather than regulars. yesterday
  • 1
    I'll also add that "many unanswered questions" is the norm on Stack Overflow because there are just that many questions.
    – F1Krazy
    yesterday
  • 1
    I don’t really know that enough time has passed since the announcement to draw any conclusion about what any high-rep user has done since, or why. Never mind “most.” yesterday
4

I don't think one can effectively protest much in the sense that it immediately and directly contributes to a different outcome. SO is a private company and has a lot of leeway to pursue partnerships. The history of the last years shows that the current company does not attribute much importance to the part of its user base that is active on Meta. Expressing an opinion additionally to the many negative scores will probably not make the message any clearer. It's already clear enough and the company really cannot say it doesn't know what the community is thinking. In that way, registering more protest would be an inefficient use of one's time.

Short of stopping to contribute and ideally starting to contribute elsewhere I see nothing that really can be done that effectively makes a differences.

CC BY-SA 4.0
1

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .