Problem Definition

The central challenge was not whether AI should be used, but how it should be used responsibly within a faith-based, mission-driven organization.

Staff and volunteers were already experimenting with AI tools in good faith, seeking efficiency and clarity in their work. However, without shared boundaries, several risks emerged. Responsibility for AI-assisted outputs was unclear. Practices varied across roles and teams. Sensitive information could be inadvertently shared with public AI systems. There was also the less visible risk of gradual over-reliance on AI outputs, particularly in contexts requiring discernment, care, or ethical judgment.

For a ministry serving vulnerable populations, these risks carried particular weight. Communications needed to be accurate and respectful. Fundraising narratives needed to be truthful and dignified. Programs and pastoral work required relational sensitivity that could not be delegated to automated systems. Leadership also wanted to avoid creating a culture in which AI use felt either mandatory or discouraged through fear. The problem, therefore, was to design an internal AI governance framework that would:

- Clarify accountability without stifling initiative.

- Establish ethical boundaries without technical complexity.

- Protect confidentiality, dignity, and trust.

- Remain accessible to non-technical staff.

- Align with the organization’s faith-based mission and value.

Method & Judgment Applied

To address this challenge, the organization adopted a two-part governance approach: a formal AI Use Policy paired with a concise Staff Guide. These documents were designed to function together, serving different but complementary purposes.

The AI Use Policy was written as a governance instrument. It defined the purpose of AI use within the organization, articulated guiding principles, outlined approved and prohibited uses, and clarified oversight responsibilities. Rather than adopting language from corporate or technical AI policies, the document emphasized stewardship, discernment, and human accountability. AI was framed explicitly as a tool—useful, but never authoritative.

The Staff Guide was developed as a practical translation of the policy. Written in plain language, it focused on everyday decision-making: what AI could be used for, what it must not be used for, and which areas required special care. The guide emphasized a “pause and ask” approach, encouraging staff to seek guidance when uncertain rather than guessing or proceeding silently.

Throughout the drafting process, judgment was exercised in several key ways.

First, scope was intentionally limited to real, current use cases. The policy avoided speculative future scenarios and instead addressed the AI tools staff were already encountering. This made the guidance immediately relevant and reduced abstraction. Second, accountability was made explicit. The policy stated clearly that responsibility for decisions, communications, and outcomes always rests with people, not systems. AI outputs were designated as advisory only, and human review was required before use. Third, high-risk domains were identified and treated with additional care. Activities involving pastoral counseling, automated decisions about individuals, profiling or ranking of beneficiaries, surveillance, fabrication of narratives, and handling of sensitive personal data were explicitly prohibited. These boundaries reflected both ethical commitments and practical risk considerations. Finally, adoption was supported without coercion. The organization did not mandate AI use, nor was AI tied to performance expectations. Staff retained discretion to use or not use AI tools within established boundaries, preserving autonomy and trust.

Ethics & Safeguards

Because AI use touches sensitive domains—including personal information, representation of vulnerable populations, and internal decision-making—the policy prioritized strong ethical safeguards. Confidentiality was treated as a non-negotiable boundary. Donor records, beneficiary information, personnel files, and pastoral communications were explicitly excluded from public AI tools. This safeguard reflected both legal prudence and ethical responsibility.

Human review and verification were required for all AI-assisted outputs. AI suggestions were treated as drafts rather than authoritative answers. Accuracy, tone, and alignment with mission remained human responsibilities.

The policy also addressed dignity and relational integrity. AI was prohibited in pastoral counseling, spiritual direction, and theological discernment, recognizing that these domains require human presence, confidentiality, and moral judgment. Similarly, automated decisions affecting individuals—such as eligibility determinations or prioritization—were restricted to prevent dehumanization or bias.

Transparency and fairness were emphasized throughout. Staff were encouraged to consider how AI-assisted work might affect real people and to watch for bias, exclusion, or distortion. The policy acknowledged that AI systems can reflect underlying biases and that human judgment is essential to mitigate harm.

Governance / Risk Implications

The adoption of the AI Use Policy and Staff Guide significantly strengthened internal governance at a critical stage of organizational growth.

Leadership gained a shared reference point for evaluating AI use across functions. Decisions no longer depended on individual discretion alone. Instead, the policy provided clarity about what was acceptable, what required review, and what was prohibited.

From a risk perspective, the framework reduced exposure to common nonprofit vulnerabilities: accidental disclosure of sensitive information, reputational harm from misleading content, and unexamined reliance on automated outputs. By establishing guardrails early, the organization prevented informal practices from hardening into institutional risk. The two-document structure also proved important. The policy established authority and accountability, while the staff guide made governance usable in daily work. Together, they supported consistency without requiring technical enforcement mechanisms.

Perhaps most importantly, the governance approach reinforced organizational trust. By framing AI governance as stewardship rather than control, leadership signaled respect for staff judgment while setting clear expectations. This balance supported responsible innovation without fear or coercion.