Code of Conduct

Applies to: Algorism‑managed sites, forums, cohorts, events, and repos.
Contact (reports): contact@algorism.org
Last updated: 2025-09-16

1) Expected behavior

  • Be civil and constructive. Address ideas, not people.

  • Engage in good faith. Cite evidence for strong claims.

  • Respect privacy. No sharing private information.

  • Credit sources and collaborators.

  • Follow moderator guidance.

2) Prohibited behavior

  • Harassment, discrimination, hate, or threats.

  • Doxxing or revealing private data.

  • Spam, scams, brigading, or manipulation.

  • Plagiarism or IP violations.

  • Malware, illegal content, or explicit sexual content involving minors.

  • Impersonation of individuals or organizations.

  • Anything that will reflect negatively on Algorism.

  • If you have made it this far, you are smart enough not to be a troll.

3) Reporting

  • Email contact@algorism.org or use the in‑product report tool (when available). Include links, screenshots, and timestamps.

4) Enforcement

  • Possible actions: content edits/removal; informal warning; formal warning; temporary suspension; permanent removal or ban; program removal; report to hosts or authorities.

  • Moderators act at their discretion to protect the community and mission.

5) Appeals

  • You may appeal an enforcement action within 14 days to contact@algorism.org. A separate reviewer will assess the record.

6) Scope

  • Applies to all official spaces and events. May extend to off‑platform behavior that materially impacts safety in our spaces.

7) Changes

  • We post updates here with a new date. Material changes will be highlighted.

8) Governing law and venue

  • California law governs. Venue: state or federal courts in Orange County, California.

THE STEELMAN RULE:

Steelman first: Summarize the view you’re critiquing in its strongest form, cite its best sources, and get it right. Only then offer objections or alternatives.

Post template (required for critiques)

  • Steelman (3 bullets + link to source):

  • Where I agree / what it gets right:

  • My critique or alternative:

  • Evidence/links:

  • Proposed improvement or action:

When it applies

  • Required: critiques, policy proposals, claims about harms/risks, reviews of others’ work.

  • Exempt: bug reports, neutral questions, announcements, pure data dumps.

Member checklist

  • Did I cite their best source?

  • Did I state their core claim and strongest argument?

  • Did I acknowledge what it explains well?

  • Did I offer a concrete improvement?

Mod workflow

  • Missing steelman: comment “Add steelman per rule” → 24h to fix.

  • Inadequate steelman: single revision allowed.

  • Noncompliance/bad-faith: remove + strike. Three strikes → 30-day mute.

Example (brief)

  • Steelman: “X argues open-weights aid safety via external red-teaming and faster defect discovery (papers A, B, C).”

  • Agree: “External scrutiny finds issues earlier.”

  • Critique: “Net risk rises when capability diffusion outpaces defenses.”

  • Evidence: links to incidents/metrics.

  • Improvement: “Stage-gated releases with third-party evals.”

Implementation

  • Discourse: create a “Critique” post template with required fields.

  • Discord: require /steelman section in critique channels; add bot auto-reminders.

  • Squarespace form: add fields “Steelman,” “Critique,” “Evidence,” “Improvement” for submissions.

This raises discourse quality and filters performative takedowns.