StackOverflow is banning accounts that delete answers in protest against OpenAI
Stack Overflow’s new partnership with OpenAI has prompted some contributors to overwrite or delete their highly rated answers in protest, leading the site to suspend accounts for “defacing” content. Commenters debate whether users should be able to retract their work, how GDPR and Creative Commons licenses apply, and whether AI training on community-generated knowledge is ethical when the resulting models are proprietary. Many see the move as part of a broader shift in which user-created content is systematically monetized by large platforms and AI companies, raising questions about ownership, consent, and the future of open knowledge.
Scope of the Bans and Policy Context
- Reports describe users replacing high‑value answers with protest text and receiving suspensions.
- Multiple commenters note that SO has long blocked deletion/defacement of popular answers to preserve the public knowledge base; this is not a new policy.
- Suspensions mentioned are temporary (e.g., 7 days), and permanent bans are said to be rare and reserved for persistent abuse.
- Some view the HN/OSNews framing as overstating the severity (“banning accounts”) versus enforcing existing rules.
Legal Rights: GDPR, Licenses, and Moral Rights
- Many argue GDPR mainly covers personal data (identifiers), not voluntarily published Q&A content.
- Others suggest stylometry might make text inherently personal, but this is contested.
- Stack Overflow content is under CC BY‑SA; commenters stress:
- Authors granted a perpetual, non‑revocable license.
- SO and anyone else can keep and reuse answers with attribution and share‑alike.
- Some mention jurisdictions with inalienable “moral rights” and speculate (unclearly) whether AI training could violate these.
OpenAI Partnership and Monetization Ethics
- Some see the deal as a natural way for SO and OpenAI to monetize shared knowledge; users already benefited from free answers.
- Others resent “the many” creating value that “the few” now monetize, especially for a closed, non‑share‑alike model.
- A notable faction says using CC‑licensed answers to train a proprietary model likely ignores attribution/share‑alike obligations; the legal status of LLM training is noted as unresolved.
Protest Tactics and Their Impact
- Ideas floated: direct defacement, gradual corruption (minor edits, homographs), or mass edits to overload moderation.
- Critics argue this harms human learners and is easy for SO to detect/roll back from archives.
- Some propose GDPR deletion requests as a more legitimate protest route, though others doubt they apply to Q&A content.
Stack Overflow’s Future and Centralization
- Opinions split: some see “death pangs” as users move to ChatGPT/Copilot; others say SO will adapt or be replaced by new tools if it fails.
- Several highlight the systemic issue: centralized platforms and public content are inherently monetizable and scrappable, partnership or not.