Stack Overflow simply bans folks who don't want their advice used to train AI
- Reference: 1715292250
- News link: https://www.theregister.co.uk/2024/05/09/stack_overflow_banning_users_who/
- Source link:
The outcry comes after it was [1]announced earlier this week that OpenAI would soon be able to train AI models using the answers and knowledge Stack Overflow users have contributed in the past 15 years. Stack Overflow and OpenAI will both benefit from the deal, with SO getting access to OpenAI technology, and OpenAI by having all that juicy data to stuff into its neural networks.
One of the Stack Overflow posters unhappy about having their data ingested without explicit permission is Ben Ui, a software developer who said he chose to protest the move by overwriting previous posts to complain about the deal. The protest landed him a week-long suspension and his posts reverted back to as they were, Ui claims.
[2]
"Stack Overflow does not let you delete questions that have accepted answers and many upvotes because it would remove knowledge from the community. So instead I changed my highest-rated answers to a protest message," Ben [3]said in a post on Mastodon. "Within an hour mods had changed the questions back and suspended my account for seven days."
[4]
[5]
Other users say they've had similar experiences, and that Stack Overflow has been banning users wholesale who have attempted to delete or deface their own posts on the site. We've asked Stack Overflow to clarify the extent of the protests, and haven't heard back yet.
Regardless, Stack Overflow's terms-of-use don't appear to give users much wiggle room, in that it appears the data can be used for AI training.
[6]
Ben shared an email he said was sent by the Stack Overflow moderation team that explained he was suspended because the stuff he posted on Stack Overflow was no longer his once it had been posted.
"Once you post a question or answer to this site, those posts become part of the collective efforts of others who have contributed to that content," Team Overflow said in the email. "Because of the amount of disruption these incidents can cause, we have placed your account on hold for seven days."
As others pointed out to those upset over Stack Overflow's response, it's all covered right there in the terms of service.
[7]
"You agree that any and all content … that you provide … is perpetually and irrevocably licensed to Stack Overflow on a worldwide, royalty-free, non-exclusive basis," Stack Overflow's [8]Subscriber Content section in its terms states.
"You grant Stack Overflow the perpetual and irrevocable right and license to access, use, process, copy, distribute, export, display and to commercially exploit such Subscriber Content, even if such Subscriber Content has been contributed and subsequently removed by you," the terms continue.
"This means that you cannot revoke permission for Stack Overflow to publish, distribute, store and use such content," the terms sum up. In short, sorry: Stack Overflow's content is all Stack Overflow content.
[9]Stack Overflow to charge LLM developers for access to its coding content
[10]Stack Overflow bans ChatGPT as 'substantially harmful' for coding issues
[11]Funnily enough, AI models must follow privacy law – including right to be forgotten
[12]The truth about Dropbox opening up your files to AI – and the loss of trust in tech
Ben, for his part, said he's requesting his Stack Overflow data be deleted under the European Union's GDPR's right to be forgotten. Though as [13]GDPR Article 17 states, the right to have your data erased doesn't apply when the processing of said data is necessary "for exercising the right of freedom of expression and information," and Stack Overflow could perhaps make the information argument considering its view of ownership of the site's content.
Wherever the fight ends up, one can imagine the lawyers quivering with anticipation at the possibility that the case lands in their territory.
In the meantime, it doesn't appear Stack Overflow users have any way to opt out of having their data used to train AI models. We've asked about that as well, and still have yet to hear back. ®
Get our [14]Tech Resources
[1] https://www.theregister.com/2024/05/07/stack_overflow_openai/
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2Zj2b4bZ@8a0EFENClF@BdwAAAE0&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://m.benui.ca/@ben/112396505994216742
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Zj2b4bZ@8a0EFENClF@BdwAAAE0&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[5] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Zj2b4bZ@8a0EFENClF@BdwAAAE0&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[6] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44Zj2b4bZ@8a0EFENClF@BdwAAAE0&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[7] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33Zj2b4bZ@8a0EFENClF@BdwAAAE0&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[8] https://stackoverflow.com/legal/terms-of-service/public#licensing:~:text=property%20rights%20therein.-,Subscriber%20Content,-You%20agree%20that
[9] https://www.theregister.com/2024/03/01/stack_overflow_launches_api_to/
[10] https://www.theregister.com/2022/12/05/stack_overflow_bans_chatgpt/
[11] https://www.theregister.com/2023/07/13/ai_models_forgotten_data/
[12] https://www.theregister.com/2023/12/15/dropbox_ai_training/
[13] https://gdpr-info.eu/art-17-gdpr/
[14] https://whitepapers.theregister.com/
# SO posts sliding down a slope, then falling off a cliff
researchgate dot net A-Time-series-of-weekly-posts-to-Stack-Overflow-since-early-2016-The-number-of-weekly_fig1_372404443
There are a multiple factors to unwrap. One is that the SE answer space was already getting saturated by 2016. The other is AI, improved copilot, etc.
I was surprised to see, last year, that CEO AK's salary was 500K - pretty low for a CEO.
Wake up kids
By now surely everyone remotely awake assumes that every word they tap into a web site is immediately and irrevocably out of their control.
Think before you type and, sadly, don't ever assume that web site owner will behave with honour.
Why, even at The Register there been cases wher.. x.&*/......
no carrier
Re: Wake up kids
If Skynet is being trained using Reg comments, or - god help us - 4chan, then we’re all well and truly boned.
Is it time to just get out? Past time?
Well if you had any residual trust in online services, this is a wake up call.
I remember when StackOverflow was in Beta, it was great.
Then there was that statement by one of the developers that they might open source it. The prevarication made it obvious, a lie.
Then you realised the game was rigged. You started an item, shortly changed your mind, couldn't delete it. Somebody had inserted a single irrelevant comment. Yep you realised, the design is pretty pure evil. Automated theft of your contribution.
After that there was a wave of "brownie point hunters" looking for upvotes irrespective that they had no idea what the question meant. (What kind of person is turned on by silly little badges? Dumb scores?)
Then came the organised gangs. I saw wrong answers top voted and right answers downvoted and "amateur lawyered" to death.
The huge numbers of outdated, and now wrong, answers to questions. The good questions unanswered. ...
Then it got sold out to enemies of us all.
Enshittification, upon enshittification. The course to the heat death of the Internet Universe. He who can't see it deserves what he'll get.
End of an era, I guess.
To be honest, SO was never good. It was only good for C# (the C# responses are invaluable and extremely in-depth, written by actual geniuses), every other language's community is either horrible and completely unhelpful, (Python answers are all written by children, every JS answer is "Just Use JQuery™", C++ users are violent animals who would rather burn you at the stake for misunderstanding their glorious language, etc.) or you usually flat-out don't need them (like how virtually everything in Rust is uber-documented, therefor SO questions/answers are mostly redundant and just say what the docs say). Maybe AI is going to raise a generation of people who deliberately do the wrong thing out of misinformation and hallucinations, but SO was never too helpful to begin with, either.
Talk about wanting all information to be free
...literally.
Just as well that I really stopped contributing years ago, after the whole Monica fiasco.