Codeberg Votes to Ban AI-Centric Projects and AI Training on User Data
Codeberg's member vote formalizes a strict anti-AI stance, banning AI-generated projects and AI training on user data, citing resource costs and trust concerns.
Reporting from 1 source: GIGAZINE.
Codeberg, the non-profit open-source hosting service, held a member vote on two AI-related proposals. Both passed by a majority. The service will not use project or user data for training AI models, and its terms of service now ban vibe-coded projects primarily created with generative AI.
Codeberg's governing body has voted to take a hard line on generative AI. The first proposal, which passed by a majority, confirms that Codeberg and its related services will not use project or user code or data for training AI tools such as LLMs. The second amends the terms of service to ban vibe-coded projects, defined as those primarily composed of code output by generative AI like Claude or OpenAI Codex.
Codeberg cited the heavy load from AI company crawlers, which generate costly database queries and degrade service quality. It also criticized single-developer projects that consume resources comparable to large community projects, calling them ghost projects that donations should not fund. The vote follows a 14-day voting period on the two proposals.
Synthesized by Yomimono from the 1 cited source below, including Japanese-language reporting where cited, then editorially reviewed before publishing.