A deal between Stack Overflow and OpenAI appears to have triggered a battle between the developer discussion board and its customers.
On Monday, Stack Overflow introduced a brand new deal wherein consumer content material could be scooped up by OpenAI to coach ChatGPT. As a discussion board for builders and programmers, Stack Overflow is residence to technical posts and content material that’s invaluable to a generative AI service like OpenAI’s ChatGPT.
The announcement, nevertheless, compelled not less than one consumer to change his posts in protest. That in flip prompted Stack Overflow to droop that consumer’s account for every week. In a put up on Mastodon, an Epic Video games developer named Ben stated that he tried to take away content material from his Stack Overflow posts in response to the discussion board’s partnership with OpenAI.
As a result of Stack Overflow does not allow you to delete questions which have accepted solutions and plenty of upvotes, Ben protested by altering the content material of his inquiries to say: “I’ve eliminated this query in protest of Stack Overflow’s determination to companion with OpenAI. This transfer steals the labor of everybody who contributed to Stack Overflow, with no method to opt-out. OpenAI has a historical past of flooding the net with inaccurate data and explicitly states that they’ll by no means pay creators for his or her work.”
Inside an hour, moderators on the discussion board had reverted Ben’s inquiries to their authentic states and suspended his account for seven days. Per Ben’s Mastodon put up, a discover from the Stack Overflow moderation crew advised him that he had just lately eliminated or defaced content material from one in all his posts.
“Please observe that when you put up a query or reply to this website, these posts develop into a part of the collective effort of others who’ve additionally contributed to that content material,” the discover added. “Posts which might be probably helpful to others shouldn’t be eliminated besides underneath extraordinary circumstances.”
The moderators concluded by saying that when the matter is resolved, his popularity rating shall be restored and his account will resume as regular.
Following up on his preliminary Mastodon put up, Ben stated that he had requested Stack Overflow to completely delete his questions and solutions underneath GDPR. He additionally criticized the OpenAI’s information scraping.
“It is only a reminder that something you put up on any of those platforms can and shall be used for revenue,” Ben wrote. “It is only a matter of time till all of your messages on Discord, Twitter, and so on. are scraped, fed right into a mannequin and offered again to you.”
Different Stack Overflow customers have chimed in with questions and complaints in regards to the OpenAI deal. One particular person requested: “The place is the opt-out choice, so my solutions do not get utilized by OpenAI?” One other particular person requested if Stack Overflow customers are legally entitled to any advantages from the OpenAI deal.”
The battle raises the query: Who owns the information that you just put up on a public discussion board? GDPR has a “proper to be forgotten” measure in which you’ll be able to request that your information be faraway from an internet site, however this sometimes applies to non-public or delicate data.
Stack Overflow’s Phrases of Service state that “when you place content material within the public sphere, you willingly surrender some rights and management over such content material.” Moreover, the discussion board helps you to delete a query or put up, however provided that nobody else has responded to it. As soon as a put up generates solutions and even upvotes, you are strongly discouraged from eradicating it.
This conflict additionally highlights points round generative AI and information gathering. What occurs when a public web site permits its content material to be scraped by AI? Do the creators of that content material have any say within the matter?
Stack Overflow is dealing with a backlash partially as a result of the corporate had beforehand resisted the lure of AI. In a coverage put up from late 2022, the location banned the usage of ChatGPT and different generative AI instruments when writing or rewriting content material. Now, AI appears to be okay, so long as it generates income for the corporate.