Codeberg Updates Terms of Service to Ban AI-Generated Code

Codeberg Bans Generative AI Code to Protect FLOSS Commons

Codeberg has officially updated its Terms of Service (TOS) to prohibit the hosting of projects that consist primarily of code written by generative AI tools. This policy change is designed to safeguard the Free and Libre Open Source Software (FLOSS) ecosystem from the proliferation of "AI slop," copyright uncertainties, and the increased maintenance burden placed on human developers by automated contributions.

New Terms of Service Requirements

The specific amendment to the Codeberg TOS explicitly forbids the sharing of projects that are mostly AI-generated. The policy states:

"You must not share projects that mostly consist of code written by 'generative AI'-tools (including services such as Claude, OpenAI Codex). Such projects having an unclear copyright status... and furthermore have little safeguards to ensure that they do not include harmful code."

This restriction is grounded in two primary concerns: the legal ambiguity surrounding the copyright of AI-generated works and the security risks associated with code produced by LLMs that may lack rigorous human oversight.

The Rationale: Community Preservation vs. Automation

Codeberg's stance is part of a broader effort to maintain a community hub for developers who prioritize human-centric software development. The platform argues that LLMs can erode the quality of the commons by turning coders into "paper pushers" and overwhelming maintainers with low-effort pull requests.

Community reactions to this move are deeply polarized:

• Support for Principled Stances: Some users view Codeberg as a vital sanctuary for those hostile to LLM agents. Proponents argue that not every platform needs to be universal and that having a dedicated space for anti-AI development provides a strong signal that a project is "not ai-slop" and is maintained by humans who care about the code.
• Concerns Over Enforcement: Critics question how Codeberg will practically detect AI-generated code. Some argue that there is no reliable way to distinguish between a human who writes clean code and an LLM, leading to fears of spurious accusations or "trolling" used to remove projects.
• Productivity Gains: Some developers argue that AI has significantly increased their participation in FOSS by eliminating tedious tasks like CI/CD configuration and packaging, claiming that the anti-AI sentiment is "academic and out-of-touch."

Governance and Platform Identity

This policy change was the result of a democratic vote by Codeberg's active members, reflecting a shift in the platform's identity from a general FLOSS alternative to GitHub toward a more specific, ideologically driven community.

While some users appreciate this democratic approach, others express concern over the stability of such governance. Critics suggest that a platform that changes its fundamental rules based on current sentiment may be unreliable for long-term hosting. However, defenders of the move argue that the ability to fork projects and host one's own Git instance is the ultimate safeguard of the FOSS philosophy, allowing different platforms to maintain distinct ideological boundaries.

Impact on Hobbyists and Small Projects

There is ongoing uncertainty regarding how this policy affects small-scale or "one-off" hobby projects. While the TOS targets projects that mostly consist of AI code, some solo developers have questioned whether the lack of a formal community around their hand-written projects makes them more vulnerable to scrutiny or if the platform is shifting its focus exclusively toward large, community-driven projects.