Akashic Engine Governance Policy

The Akashic Engine is a community-governed feed intelligence system that allows users to choose between two content ranking modes: Human Mode (engagement-based ranking) and AI Mode (Cloudflare Workers AI recommendations guided by a collectively authored positivity prompt).

This Governance Policy describes how the community participates in shaping the AI Mode positivity prompt, the rules governing proposals and voting, and the rights and responsibilities of all participants.

Human Mode: Content is ranked purely by engagement signals. Posts are scored by sparks (reactions) weighted by 2 plus comment count. Videos are scored by view count. No AI influence is applied. Rankings are deterministic and verifiable.

AI Mode: Content recommendations are generated by Cloudflare Workers AI with the active community-authored positivity prompt injected as contextual guidance. The prompt shapes recommendation behavior but does not override safety filters or content policies.

Users may switch between modes at any time. Mode preference is stored per-user and does not affect other users' experiences.

The positivity prompt is the textual instruction that guides AI Mode recommendations. It is authored, versioned, and governed by the community through the following democratic pipeline:

• Propose: Any authenticated user in good standing may submit a new prompt proposal. Proposals must include the full proposed prompt text and a rationale for the change.
• Vote: Community members may cast one vote (for or against) per proposal. Votes cannot be changed once cast.
• Quorum: When a proposal reaches the required vote threshold, voting automatically closes and the proposal advances to review.
• Admin Review: Platform administrators review proposals that pass quorum. Admins may approve or reject with written notes explaining their decision.
• Activation: Approved proposals become the new active prompt. The previous prompt is archived with full version history preserved.

Users participating in the governance process must:

• Maintain an account in good standing (not suspended or restricted)
• Submit proposals that comply with Community Guidelines
• Not engage in vote manipulation, sockpuppet voting, or coordinated inauthentic behavior
• Respect the democratic process and accept outcomes

Proposals that contain harmful content, violate Community Guidelines, attempt to manipulate recommendations for commercial gain, or undermine platform safety will be rejected regardless of vote count.

Proposed prompts must:

• Promote positive, constructive, and inclusive content discovery
• Not target, demote, or suppress content from specific individuals or groups
• Not attempt to circumvent content safety filters
• Not include personally identifiable information
• Not promote illegal activity, violence, or discrimination
• Be written in clear, understandable language
• Not exceed 2,000 characters in length

All prompt versions are permanently recorded with timestamps and version numbers. The full prompt history is publicly visible to all authenticated users. Vote counts (for/against) are visible on each proposal. Admin review decisions and notes are recorded and visible.

We believe algorithmic transparency is a fundamental right. The Akashic Engine exists to give users meaningful choice and collective agency over how AI shapes their content experience.

The following data is collected and stored in connection with the Akashic Engine:

• Mode preference: Which feed mode (Human/AI) a user has selected. Stored as a user setting.
• Proposal submissions: Proposal text, rationale, and author identity are stored and publicly visible.
• Votes: Individual vote records (user ID, proposal ID, vote direction) are stored to enforce one-vote-per-user and calculate totals. Individual votes are not publicly displayed.

This data is processed under our Privacy Policy. Mode preferences and vote records are retained as long as the account is active.

Ascended Social reserves the right to:

• Reject any proposal that violates this policy or Community Guidelines
• Modify quorum thresholds to maintain governance quality
• Temporarily disable the governance pipeline for safety or maintenance
• Revert to a previous prompt version if an active prompt causes harm
• Modify Human Mode ranking signals with advance notice to users
• Update this Governance Policy with reasonable notice

Users who believe a governance decision was made in error may contact support. Disputes will be reviewed by the platform trust and safety team within 14 business days.

We may update this Governance Policy from time to time. Material changes will be communicated through in-app notification at least 14 days before taking effect. Continued use of the Akashic Engine after changes take effect constitutes acceptance of the updated policy.