AI Policy
Effective date: August 29, 2026
Commonway System uses AI tools throughout its research, drafting, and development, and this page says so plainly rather than leaving it unstated. AI does not have final say over anything published here. Every classification, citation, and design decision is reviewed and decided by a human before it goes live.
#What AI was used for
Claude, built by Anthropic, has been closely involved throughout this project: researching source material, drafting pattern content, structuring the classification system, building the brand guide, and assisting with the site's own code through Claude Code. Commonway System would look very different, and would have taken far longer to build, without it. We think that's worth crediting openly rather than treating as incidental.
#Human oversight
AI drafts, researches, and suggests. It doesn't decide. Every pattern's classification, every citation, every structural decision in the guide is reviewed against the same evidence-tier system described on the Governance page, legal or regulatory sources first, then evidence-based research, then documented precedent, before it's published. An AI-drafted paragraph doesn't get special trust for being AI-drafted; it gets checked the same as anything else.
#Why this matters for a citation-based project
AI systems can misstate a source, misread a number, or state something with more confidence than the underlying evidence actually supports. That's exactly the failure mode the guide's citation and evidence-tier system already exists to catch, whether the original error came from an AI, a contributor, or the maintainer's own mistake. The unsourced-claim flag⚠ applies the same way regardless of who or what drafted the claim it's attached to.
#What AI does not do here
There's no AI chatbot, voice agent, or automated assistant on this site talking to visitors. AI isn't used to auto-generate a pattern's classification or citation without a human checking it against a real source first. Nothing on this site is AI-generated content published without review.
#Limitations
AI-assisted drafting can contain errors, outdated information, or misattributed sources, the same as human drafting can. This is exactly why nothing on this site is treated as authoritative on the strength of who or what wrote it; see the Terms and Conditions page for the site's broader disclaimer on accuracy and reliance.
#Changes to this policy
This page may be updated as the tools and processes behind Commonway System change. Updates are reflected in the site's version history; see Release History & Roadmap.
#Contact
Questions about this policy can be directed to [email protected].