Stack Overflow Moderators Stop Work in Protest of Lax AI-Generated Content Guidelines (gizmodo.com) 41
Moderators of Stack Overflow have announced a strike in protest of the company's ban on moderating AI-generated content, claiming that this policy allows incorrect information and plagiarism to proliferate on the platform. Gizmodo reports: Last week in a post -- which has been downvoted at least 283 times -- Stack Overflow announced its new moderation policy that will only remove AI-generated content in specific instances, claiming that over-moderation of posts made with artificial intelligence was turning away human contributors. The company also said in its post that a strict standard of evidence needed to be used moving forward in order to manage AI content, and that that standard of evidence hasn't applied to most suspensions issued by moderators thus far. This directive was also communicated to the platform's moderation team privately before being posted publicly. The moderators of the website are claiming that this directive will allow AI content, which can frequently be incorrect, to run rampant on the forum while expressing discontent with Stack Overflow for not communicating this new policy more effectively.
"Stack Overflow, Inc. has decreed a near-total prohibition on moderating AI-generated content in the wake of a flood of such content being posted to and subsequently removed from the Stack Exchange network, tacitly allowing the proliferation of incorrect information ("hallucinations") and unfettered plagiarism on the Stack Exchange network. This poses a major threat to the integrity and trustworthiness of the platform and its content," the mods write in their letter to Stack Overflow. "Stack Overflow, Inc. has decreed a near-total prohibition on moderating AI-generated content in the wake of a flood of such content being posted to and subsequently removed from the Stack Exchange network, tacitly allowing the proliferation of incorrect information ("hallucinations") and unfettered plagiarism on the Stack Exchange network. This poses a major threat to the integrity and trustworthiness of the platform and its content," the mods write in their letter to Stack Overflow.
Stack Overflow moderators, like those at Wikipedia, are volunteers tasked with maintaining the integrity of the platform. The moderators say that they tried to express their concerns with the company's new policy through proper channels, but their anxieties fell on deaf ears. The mods plan to strike indefinitely, and will cease all actions including closing posts, deleting posts, flagging answers, and other tasks that help with website upkeep until AI policy has been retracted.
"Stack Overflow, Inc. has decreed a near-total prohibition on moderating AI-generated content in the wake of a flood of such content being posted to and subsequently removed from the Stack Exchange network, tacitly allowing the proliferation of incorrect information ("hallucinations") and unfettered plagiarism on the Stack Exchange network. This poses a major threat to the integrity and trustworthiness of the platform and its content," the mods write in their letter to Stack Overflow. "Stack Overflow, Inc. has decreed a near-total prohibition on moderating AI-generated content in the wake of a flood of such content being posted to and subsequently removed from the Stack Exchange network, tacitly allowing the proliferation of incorrect information ("hallucinations") and unfettered plagiarism on the Stack Exchange network. This poses a major threat to the integrity and trustworthiness of the platform and its content," the mods write in their letter to Stack Overflow.
Stack Overflow moderators, like those at Wikipedia, are volunteers tasked with maintaining the integrity of the platform. The moderators say that they tried to express their concerns with the company's new policy through proper channels, but their anxieties fell on deaf ears. The mods plan to strike indefinitely, and will cease all actions including closing posts, deleting posts, flagging answers, and other tasks that help with website upkeep until AI policy has been retracted.
They're playing a dangerous game (Score:2)
What if the site improves overall? At least some of the moderators on there seem more like self-important martinets than anything else - the middle managers of the Stack Overflow world. It's a reasonable question to consider whether the loss of those people's "service" will result in a net loss or gain in quality of the site.
Re: (Score:2)
I was just thinking that for one thing people might be able to ask legitimate questions again.
Re: (Score:2)
More than one SE site has mods which actively suppress other answerers because they prefer to get the karma themselves - seen this a few times now.
Re: (Score:3, Informative)
Re: They're playing a dangerous game (Score:5, Insightful)
The purpose of stack overflow is to train AI? That doesn't sound right.
Re: (Score:2)
It was back when the FOTM "AI" in question were "Expert Systems", before the needle moved.
Re: (Score:1)
People giving unpaid content to a company that uses their answers to train AI and make money.
Is this like a self-esteem crisis or something? Why are people working for free for commercial enterprises, and creating free produce for a company?
SE must be flighty stinking rich. They do not pay for anything. What a brilliant business model! I just cannot believe that people do it.
Re:They're playing a dangerous game (Score:5, Interesting)
If people wanted answers from ChatGPT, they'd go to chat.openai.com to get them. OTOH if they are at Stackoverflow.com, they expect the answers to be served up by people, because that is what the site is about.
People who copy-and-paste ChatGPT's content onto StackOverflow just waste everybody's time.
Sounds like this could be a good thing... (Score:2)
Re:Better the devil you know... (Score:4, Interesting)
Re: (Score:2)
The problem with Stack Overflow and dupes is that:
1. The manual may be incomplete, or outdated, or inapplicable to the situation being asked by the questioner.
2. The questioner has to take the time and effort to research every last permutation of the question being asked and explain in the post why it isn't applicable to this situation.
3. And sometimes, in spite of all this, the question is still voted as a dupe because people don't read the post (ironically).
To make things worse, there are people out there [stackoverflow.com]
Re: (Score:1)
LOL, you get reputation, which you cannot spend anywhere or buy anything with, by creating and giving content away to a company, who turns around and makes money on it.
Um, good job. I guess.
My God, these people are brilliant.
I am going to start paying people in CoolNuggets from now on. Apparently people WILL accept them as payment, and they cost me nothing.
Cha-motherfucking-Ching!
Re: (Score:1)
BTW, my car is absolutely filthy. I am willing to offer 100 CoolNuggets to whoever washes it for me.
Awww hell, I am feeling generous, make it 1,000 CoolNuggets.
Consider this the sign-up sheet.
Re: (Score:1)
I hear you can increase your penis size by taking this little pill every day. It's made from some deep jungle unknown plant that only the ancient Inca people knew about.
They stopped (Score:1)
If it's labelled then what's the issue? (Score:1)
Someones doing something, however I think it will be too little to late [slashdot.org]
Really? (Score:4)
Perhaps it's just the tags that I'm subscribed to. But I haven't noticed a damn's worth of difference today. Stackoverflow is still the same dumpster fire it's been for the last couple of years.
(just checked -- nope, it's still an endless stream of wandering, confused souls, who keep getting confused between stackoverflow.com, and pleasedomyhomeworkforme.com, writemycodeforme.net, or pleasesolvemycodingpuzzleforme.org)
'All bad answers should be respected' (Score:2)
As far as I can tell Stack Exchange's reasoning seems to be that they can't tell bad AI generated spam answers apart from answers where offshore outsourcers shotgun copy and paste seemingly related but wrong answers in an effort to get some upvotes for their resume.
So since they can't tell the two type of obviously wrong answers apart they might as well allow them both because that means more hits for Stack Exchange.
Obviously I can see why the moderators would be unhappy with that.
Re:'All bad answers should be respected' (Score:4, Insightful)
Re: (Score:2)
The parent is making as guess about what the real reasoning could be. I don't know how plausible that is, I didn't know that internet points were something that you could put on your resume, but that was the mean
Re: (Score:2)
what is the reasoning behind limiting the moderation of those bad answers?
-Too many knee-jerk moderations of "looks like AI generated text, you are banned" instead of properly reviewing the content posted. Applying a ban-hammer too heavily drives away members and costs the site revenue.
-New rule: Stop banning posters, and moderate posts based on content. Site makes more money.
Re: (Score:2)
Re: (Score:2)
> too many knee-jerk moderations of "looks like AI generated text, you are banned"
Yeah, that's a made-up problem. Here's actual statistics from the Physics Meta:
https://physics.meta.stackexchange.com/questions/14438/physics-se-remains-a-site-by-humans-for-humans?cb=1
As you can see, these "knee-jerk" mods have *messaged* a grand total of 16 people, 10 of which were accounts that were created, posted bogus answers all in one day, and then never returned.
Yeah, totally "knee-jerk".
Re: (Score:2)
> that would risk drastically lowering noise to signal ratio of SO
The lower the noise-to-signal ratio, the better. Ideally zero.
Re: (Score:2)
Well in the Physics sub it's not so hard to find the wrong answers.
"how many neutrons are in a liter of water" - note, you have to use US spelling or it doesn't even parse it
Bing: there are no neutrons in a liter of water
CGTP3.5: 556
CGTP4.: 2x the correct number
One might be inclined to say "look, it's getting better!", but one will never be able to trust these things.
That might be addressed if they simply put up their numbers on the weightings. Little pop-ups that say things like "I'm 85% confident in this
wait, what? (Score:2)
Wrong things on Stack Exchange?
No! That could never happen!
Over moderation issues? (Score:3)
claiming that over-moderation of posts made with artificial intelligence was turning away human contributors
LMFAO, when has SO ever cared about over moderation? Their moderation policies have kept people who don't care about earning "reputation" on the site away for ages, and kept lots of people from asking good questions, and having tons of good and unique questions closed as duplicates.
Re: (Score:1)
A moderator strike might actually improve SO a lot. The only sub-sites that actually work are the ones where nobody votes or moderates anything.
It's not about moderation (Score:5, Insightful)
Crap overflow (Score:2)
That is essentially what will happen if they continue to let the crap AI "writes" in.
Ok (Score:3)
The Earth is flat and everyone is a filthy dirty monster unworthy of living because they don't believe the same stuff as I. Also the Stackoverflow CEO eats cooked babies.
-This post is AI generated so you are not allowed to moderate it.-
New Tulip Mania (Score:3)
Now that crypto bored masteurbating apes yacht club NFT ABC XYZ whatever is sunsetting AI hype is taking place as the latest tulip mania. And this mania is starting to show it's nasty seedy side like what happened with cryptocurrency.
So this means more hard work for people who are trying to seperate the reality from the bullshit
Here we go again...
Was a good site... (Score:2)
StackOverflow was incredible, when it started. But cruft accumulates - no one wants out-of-date answers to current problems. Then students started posting their homework questions.
So now it's going to be filled with ChatGPT hallucinations. Why? Who is posting that? If I wanted ChatGPT's opinion, I would ask it directly.
Assanine claim (Score:2)
"over-moderation of posts made with artificial intelligence was turning away human contributors."
I expect the reality is exactly the opposite, because who wants to compete with bots.
Sorry. (Score:1)
These mods (Score:2)
The Biggest Surprise (Score:1)
Is that there are people working for free, for a billion dollar company.
Why would anyone work for fee, when others are making money from their work?