Ethical Considerations in AI Grantmaking

Ethical AI keeps people in control, protects privacy and ensures decisions are explainable and fair.

• Human‑in‑the‑loop for all approvals.
• Clear statements of AI use and data handling.
• Monitoring for bias and error with corrective action.

Governance essentials

Set policy before deployment.

• Scope of AI use and exclusions.
• Logs of prompts, outputs and edits.
• DPIAs and incident response routes.

Key takeaway: policy and practice must align.

Transparency and accountability

Be open with applicants and boards.

• Explain what AI does and does not do.
• Provide channels for questions and corrections.
• Publish fairness and performance summaries.

Key takeaway: Plinth supports explainable, auditable workflows.

Inclusion and equity

Design to reduce burden and bias.

• Accessible processes and flexible formats.
• Regular review of outcomes across groups.
• Support for small or minority‑led organisations.

Key takeaway: ethics show up in everyday design choices.

FAQs

Can AI make final decisions?

No—keep humans responsible.

What data can we use?

Use only what is lawful and necessary; avoid sensitive data where possible.

How do we handle mistakes?

Record, correct and learn; update prompts and processes.