Every AI policy should start with the question: What values are we protecting? Organizations should frame AI as a tool, not a replacement—one that supports mission-driven work without diminishing the human contribution.

Pro Tip: If your organization has a certification such as 1% For The Planet, B Corp, Women Owned Business, Fairtrade, etc., use the principles from that framework to shape your AI policy – focusing on equity, sustainability, and transparency.

Example: A sustainability-focused nonprofit should create a checklist to evaluate how any AI use would align with its community impact goals and mission values.

AI systems have a bias problem. For values-led companies, it’s essential to explicitly address and mitigate this in your AI use. Equity-seeking groups, Indigenous communities, and marginalized populations must be prioritized in your ethical assessments.

Social Impact AI Policy Tip: Examine how AI-generated outputs may impact representation in visuals, messaging, or data interpretation.

Example: A climate tech firm paused all AI-generated image use until they could confirm bias mitigation in demographic prompts.

An Ethical AI Policy should reinforce that ethical use of AI requires openness. Organizations should make clear, internally and externally, when AI has supported content creation or research. No AI-generated content should be used without human review and acknowledgment.

Example: A consulting firm includes AI contribution notes in their project files and ensures every public deliverable includes human validation. (ie, “This article was drafted with 80% human lead content, 20% AI generated content.”)

AI should augment – not replace – strategic and ethical judgment. Require at least two team members to review any content that integrates AI, particularly before it’s shared externally.

Example: A mission-driven health brand added a mandatory double-review system to screen AI-assisted communications for tone, accuracy, and community sensitivity.

Strong data privacy and protection measures must be enforced. Avoid inputting sensitive client or stakeholder data into open AI tools, and only use platforms that meet your privacy standards. For example, paid enterprise tools like Microsoft Azure OpenAI or private instances of Anthropic’s Claude offer data encryption and do not retain user input. In contrast, free versions of AI platforms like ChatGPT or Google Gemini may store prompts and are not suitable for any content containing sensitive or proprietary data.

Best practice: Align your AI data practices with your existing governance and transparency commitments.

Example: A social enterprise used placeholder terms in their prompt engineering and confined prompt history to a closed, internal environment.

AI’s environmental cost is real. Organizations committed to climate action should include AI-related energy use in their sustainability tracking and offset those emissions annually.

Tip: Add AI emissions to your environmental impact reporting and consider offsetting through certified partners.

Example: One impact consultancy estimated its AI-driven emissions based on usage hours and carbon intensity, then offset through a verified tree planting initiative.

An Ethical AI Policy must evolve alongside the tools it governs. Build in regular review cycles and invite staff to share observations and ethical concerns about AI use.

Example: A foundation revised its AI policy quarterly, based on employee input and new legal standards around data use and copyright.