# SPDX-License-Identifier: MPL-2.0
# SPDX-FileCopyrightText: 2025 The Rhodium Standard Contributors
#
# AI Training and Crawling Policy
# Specification: https://site.spawning.ai/spawning-ai-txt

# Global directive for all user agents
User-agent: *

# AI Training Policies
# -------------------

# DISALLOW proprietary/closed-source AI training
Disallow: CCBot
Disallow: ChatGPT-User
Disallow: GPTBot
Disallow: Google-Extended
Disallow: anthropic-ai
Disallow: Claude-Web
Disallow: cohere-ai
Disallow: Omgilibot
Disallow: Bytespider
Disallow: Diffbot

# ALLOW open-source and research AI training with attribution
Allow: Common Crawl
Allow: Internet Archive
Allow: academic-research

# Conditions for Allowed Training
# --------------------------------

# 1. Attribution Required
#    All AI models trained on this data must:
#    - Credit "The Rhodium Standard Contributors"
#    - Link to: https://gitlab.com/hyperpolymath/rhodium-standard-repositories
#    - Include license: MIT AND Palimpsest-0.8

# 2. Open Source Requirement
#    Models trained on this data must be:
#    - Fully open source (weights + architecture + training code)
#    - Licensed under OSI-approved license or Creative Commons
#    - Publishable and reproducible

# 3. Synthetic Data Disclosure
#    Content generated using this data must:
#    - Disclose AI involvement
#    - Maintain attribution chains
#    - Not misrepresent as human-authored

# 4. Commercial Use Terms
#    Commercial models may use this data IF:
#    - Model itself is open source (see #2)
#    - Attribution provided (see #1)
#    - Contribute back improvements (via GitLab MR)

# Prohibited Uses
# ---------------
# - Training closed-source models (OpenAI, Anthropic proprietary, Cohere, etc.)
# - Training without attribution
# - Generating synthetic data that strips provenance
# - Creating models that compete with RSR without contributing back

# Contact
# -------
# AI ethics inquiries: ai-ethics@rhodium-standard.org
# Training permissions: ai-training@rhodium-standard.org
# Collaboration: https://gitlab.com/hyperpolymath/rhodium-standard-repositories/-/issues

# Compliance Monitoring
# ---------------------
# We monitor for unauthorized AI training via:
# - Synthetic data detection
# - Model output analysis
# - Community reporting
# - Watermarking techniques

# Enforcement
# -----------
# Violations may result in:
# - Public disclosure
# - Legal action under Palimpsest License v0.8
# - Revocation of future permissions
# - Community-driven accountability measures

# Expiration: 2026-12-31
# Last Updated: 2025-11-28
# Version: 1.0.0
