Is your ombuds office thinking about using artificial intelligence (AI)? Establishing strong ethical standards for the use of AI is important to maintain trust, fairness, and accountability—especially since your office serves as a neutral, independent, and confidential conflict resolution body. Here are some of the ethical standards you might want to include in your office’s AI policy:

1. Transparency

Definition: Be clear and open about how AI is used.

• Why it matters: People must understand when and how AI is involved—especially in sensitive matters like complaints or whistleblowing.
• Best practices:
• Notify individuals when AI tools are used in intake, analysis, or recommendations.
• Publish clear descriptions of AI systems’ functions and limitations.
• Maintain documentation of AI decision-making processes.

2. Human Oversight

Definition: Ensure that AI supports, but does not replace, human judgment.

• Why it matters: Ombuds roles require nuance, empathy, and discretion—qualities that AI lacks.
• Best practices:
• AI should augment, not automate, dispute analysis or case triage.
• All final decisions or interpretations must rest with a human ombuds.
• Encourage critical review of AI-generated outputs.

3. Data Privacy & Confidentiality

Definition: AI systems must protect sensitive personal and organizational data.

• Why it matters: Confidentiality is a cornerstone of ombuds work; a breach could undermine the entire office.
• Best practices:
• Use encryption and secure data storage for AI tools.
• Limit access to training and operational datasets.
• Avoid third-party AI tools that don’t meet strict privacy standards.

4. Fairness & Non-Discrimination

Definition: AI must not reinforce or exacerbate bias.

• Why it matters: Many ombuds users come from marginalized or vulnerable groups.
• Best practices:
• Audit algorithms for bias regularly (race, gender, language, etc.).
• Ensure diverse and inclusive data sets for training AI tools.
• Include stakeholders in tool development and evaluation.

5. Accountability

Definition: The ombuds office must be responsible for the effects of AI use.

• Why it matters: When AI causes harm or error, there must be a clear process for redress.
• Best practices:
• Assign internal roles for AI ethics governance.
• Establish a mechanism for reporting concerns or malfunctions.
• Create a protocol for halting AI use if it undermines ombuds value

5. Continuous Evaluation
   
   

   Definition: AI systems must be regularly reviewed and improved.

• Why it matters: Contexts change, and so does the impact of AI.

• Best practices:
• Collect feedback from users and staff.
• Include external, independent reviews of AI tools.
• Phase in new tools with pilots and adjust based on results.

In Summary:

Good ethical standards for AI use in an ombuds office should cover: