Skip to content

AI Policy

Artificial intelligence is rapidly evolving, and we believe in leveraging it for practical innovation. We welcome AI as a collaborative tool to enhance human capability, but we strictly reject the wave of low-effort "AI slop" and corporate automation hype. All contributions to RuleFlow must remain fundamentally human-centered.

To maintain the integrity, legal standing, and mathematical rigor of the project, please adhere to the following guidelines when using AI assistants in your workflow:

  • Rigorous Human Review: All AI-assisted work must be fully tested and thoroughly reviewed by a human before a pull request is opened. Low-effort, unverified AI code dumps will be outright rejected.
  • No Autonomous Agents: Code built and submitted by "no-human-in-the-loop" agents is strictly banned. You must understand, orchestrate, and take full responsibility for every line of code you submit.
  • Mathematical and Logic Verification: RuleFlow deals heavily with rigorous causal networks, multi-way branching, and vector memory boundaries. AI assistants frequently hallucinate complex logic. Code running without syntax errors is not enough; the underlying math, algorithmic efficiency, and memory topology must be explicitly human-verified.
  • Intellectual Property and Licensing: AI tools can inadvertently regurgitate copyleft (e.g., GPL) or proprietary code snippets. You bear full responsibility for ensuring all AI-generated code you submit is legally your own and fully compatible with RuleFlow's MIT license.
  • No AI Fluff in PRs or Commits: We value clear, concise project communication. Do not use AI to generate 500-word, overly enthusiastic pull request descriptions or commit messages for simple changes. Write your summaries directly and plainly.
  • Documentation Assistance: You are welcome to use AI to help draft, format, or refine documentation, tutorials, and guides, provided the final output is meticulously reviewed for accuracy, clarity, and a natural tone.
  • Disclosure Requirements: If you use AI to generate significant portions of code or written content, you must briefly disclose this in your pull request description. Minor daily assists, such as basic refactoring or simple autocompletes (e.g., inline GitHub Copilot suggestions), do not require disclosure.
  • OpenInfra Foundation - Policy for AI-Generated Content: Guidelines emphasizing human-in-the-loop development, "Generated-By" attribution, and the strict responsibility of contributors to verify security and license compatibility.
  • Apache Software Foundation - Generative Tooling Guidance: Practical rules for open-source contributors on evaluating the copyright, licensing, and security risks associated with AI-generated code.
  • ACM Policy on Authorship: The Association for Computing Machinery’s standards for acknowledging generative AI in academic research, enforcing that AI systems cannot be listed as authors and that humans bear full responsibility for mathematical and factual veracity.
  • GitHub Copilot Trust Center: Resources on secure, compliant, and transparent AI-assisted software development, including details on code-matching filters and intellectual property safeguards.