On Monday, Stack Overflow and OpenAI introduced a brand new API partnership that can combine Stack Overflow’s technical content material with OpenAI’s ChatGPT AI assistant. The deal has sparked controversy amongst Stack Overflow’s person neighborhood, with many expressing anger and protest over the usage of their contributed content material to help and practice AI fashions.
“I hate this. I am simply going to delete/deface my solutions one after the other,” wrote one person on sister website Stack Change. “I do not care if that is towards your foolish insurance policies, as a result of as this announcement exhibits, your insurance policies can change at a whim with out prior session of your stakeholders. You do not care about your customers, I do not care about you.”
Stack Overflow is a well-liked question-and-answer website for software program builders that enables customers to ask and reply technical questions associated to coding. The location has a big neighborhood of builders who contribute data and experience to assist others remedy programming issues. Over the previous decade, Stack Overflow has grow to be a closely utilized useful resource for a lot of builders looking for options to widespread coding challenges.
Beneath the introduced partnership, OpenAI will make the most of Stack Overflow’s OverflowAPI product to enhance its AI fashions utilizing content material from the Stack Overflow neighborhood—formally incorporating info that many imagine it had beforehand scraped with out a license. OpenAI will even “floor validated technical data from Stack Overflow straight into ChatGPT, giving customers easy accessibility to trusted, attributed, correct, and extremely technical data and code backed by the thousands and thousands of builders which have contributed to the Stack Overflow platform for 15 years,” in response to Stack Overflow.
In return, OpenAI plans to offer attribution to the Stack Overflow neighborhood inside ChatGPT, however how the corporate will try this precisely is unclear. Stack Overflow will even use OpenAI expertise in its growth of OverflowAI, an AI mannequin introduced in July 2023 that makes use of an LLM to offer solutions to developer questions.
Whereas the businesses tout the collaboration’s advantages, many Stack Overflow customers have expressed their displeasure with the deal. That is very true contemplating that till very lately, Stack Overflow appeared to take a adverse stance towards generative AI normally, banning solutions written utilizing ChatGPT. It was additionally extensively reported final yr that ChatGPT’s reputation had severely diminished Stack Overflow’s visitors, although the corporate appeared to later refute that, claiming defective evaluation by outsiders.
For the reason that announcement, some customers have tried to change or delete their Stack Overflow posts in protest, arguing that the transfer steals the labor of those that contributed to the platform with out a method to choose out. In retaliation, Stack Overflow employees have reportedly been banning these customers whereas erasing or reverting the protest posts. On Monday, a Stack Overflow person named Ben took to Mastodon to share his expertise of getting suspended after posting a protest message:
Stack Overflow moderators have acknowledged that after posts are made, they grow to be “a part of the collective efforts” of different contributors and will solely be eliminated below extraordinary circumstances, in response to The Verge. Stack Overflow’s phrases of service additionally state that customers can’t revoke permission for Stack Overflow to make use of their contributed content material.
Whereas Stack Overflow owns person posts, the location makes use of a Artistic Commons 4.0 license that requires attribution. We’ll see if the ChatGPT integrations, which haven’t rolled out but, will honor that license to the satisfaction of disgruntled Stack Overflow customers. For now, the battle continues.
This story initially appeared on Ars Technica.