EOL/IERB_GUIDELINES.md

Independent Ethics Review Board (IERB) Guidelines

For the Ethical Open License 1.0

---

1. Purpose

The IERB ensures ethical compliance with the Ethical Open License by:

• Interpreting and enforcing prohibited use cases (e.g., mass surveillance, discrimination).
• Reviewing alleged violations and issuing rulings.
• Publishing transparent guidelines and decisions to uphold accountability.

---

2. IERB Composition

2.1 Membership

• 7 members total are appointed by a multi‑stakeholder committee comprising representatives from international human rights organizations, diverse open‑source communities, and legal experts. This collaborative approach ensures balanced representation and reinforces board independence.
  • 2 legal experts: Specializing in international law, intellectual property, or human rights.
  • 2 open‑source advocates: Representatives from major open‑source organizations (e.g., OSI, Apache Foundation).
  • 3 domain‑specific ethicists: Experts in AI ethics, data privacy, or sociotechnical systems.
• Term: 3 years, renewable once.

2.2 Conflict of Interest

• Members must recuse themselves from cases involving their employers, close associates, or competing projects.

---

3. Prohibited Uses

For clarity and consistency, the following key terms are defined explicitly:

• Mass Surveillance: The systematic collection and monitoring of personal data without explicit consent, in violation of internationally recognized privacy standards (e.g., GDPR, UN Guiding Principles).
• Unethical Practices: Actions that violate a clearly enumerated list of human rights benchmarks, including those established by the UN Guiding Principles.

The IERB evaluates violations based on the following criteria:

3.1 Mass Surveillance

• Examples:
  • Facial recognition tools targeting marginalized groups.
  • Bulk data collection for non-transparent government tracking.

3.2 Warfare & Oppressive Regimes

• Examples:
  • Software integrated into autonomous combat drones.
  • Predictive policing tools used to suppress dissent.

3.3 Discrimination & Hate Speech

• Examples:
  • AI models trained on the Software to deny loans based on ethnicity.
  • Social media algorithms promoting extremist content.

3.4 Unethical Practices

• Definition: Use cases that violate the explicitly defined human rights benchmarks above.

---

4. Violation Review Process

4.1 Complaint Submission

• Complaints must include:
  • Evidence: e.g., code snippets, screenshots, usage documentation.
  • Impact Statement: Explanation of the harm caused.

4.2 Preliminary Assessment

• The IERB determines if the complaint is actionable within 7 days.

4.3 Formal Review

• Timeline: 60 days from acceptance.
• Steps:
  1. Notify the accused party and request a response (15 days).
  2. Gather evidence (e.g., third-party audits, expert testimony).
  3. Hold a private hearing (optional).
  4. Deliberate and issue a majority‑rules decision.

4.4 Outcomes and Enforcement

• No Violation: Case dismissed.
• Violation Confirmed:
  • The accused party must submit a remedial action plan with clearly defined, objective criteria for compliance during a 30‑day cure period.
  • Failure to meet these criteria will result in the termination of license rights.
  • A public disclosure of the decision (redacted for sensitive data) will be issued.

---

5. Appeals & Arbitration Process

• Appeals: Filed within 30 days of an IERB ruling.
• Process:
  • Binding arbitration under UNCITRAL rules.
  • The designated venue in Zürich now includes provisions for remote hearings or alternative neutral locations, especially for parties outside Europe.
  • Arbitrators: 3 neutral experts selected by the Swiss Arbitration Association.

---

6. Transparency & Reporting

• Published Materials:
  • Annual report summarizing cases and trends.
  • Redacted decisions on a public repository.
• Guideline Updates:
  • Revised every 2 years with community feedback.

---

7. Code of Conduct for IERB Members

• Impartiality: Decisions based solely on evidence, not personal bias.
• Confidentiality: Protect sensitive information disclosed during reviews.
• Accountability: Members violating ethics rules may be removed by the Software Freedom Conservancy.

---

Appendix A: Examples of Ethical vs. Unethical Use

Ethical	Unethical
Healthcare diagnostics for underserved communities	Predictive policing tools targeting minorities
Climate change modeling	Deepfake tools for political manipulation

---

Approved by the IERB on [DATE]
These guidelines are supplemental to the Ethical Open License 1.0 and enforceable under its terms.