Artificial intelligence (AI) is changing how organizations work, offering new opportunities to organize information, analyze data, communicate findings, and reduce administrative burden. But as these tools become more accessible, organizations must also consider how to use them responsibly, particularly when working with sensitive information, community partners, and evaluation findings that inform important decisions.
The Responsible AI Use in Evaluation Guide, last updated in June 2026, helps nonprofit and social impact organizations understand how AI can support evaluation activities while maintaining the ethical standards, accuracy, transparency, and accountability essential to high-quality work. Drawing on BSRI’s commitment to rigorous, people-centered evaluation, the guide provides practical recommendations for integrating AI into everyday workflows without compromising privacy, equity, or community trust.
Rather than presenting AI as a replacement for traditional evaluation practices, the resource encourages organizations to approach it as a supportive tool. AI can help teams work more efficiently, but human expertise remains essential for interpreting findings, understanding community context, identifying limitations, and making informed decisions.
The guide introduces five core principles for responsible AI use:
- Human Oversight and Accountability: Maintain human responsibility for reviewing, interpreting, and validating AI-generated outputs.
- Privacy and Confidentiality: Protect sensitive participant information and follow applicable privacy regulations, organizational policies, and data security requirements.
- Accuracy and Verification: Review AI-generated information for errors, incomplete findings, and inappropriate interpretations before using it in evaluation products.
- Equity and Bias Awareness: Recognize the potential for AI to reinforce biases, overlook community perspectives, or generate content that does not reflect the populations being served.
- Transparency: Communicate when and how AI contributes to evaluation activities, particularly in data analysis, reporting, and other products shared with stakeholders.
The resource also explores practical applications of AI throughout the evaluation process, including:
- Organizing and preparing datasets for analysis.
- Summarizing qualitative information and identifying recurring themes.
- Supporting report writing, executive summaries, and communication materials.
- Brainstorming survey questions, interview prompts, and evaluation tools.
- Assisting with literature reviews and research summaries.
- Streamlining administrative activities such as meeting notes, agendas, and project documentation.
Importantly, the guide addresses the risks and limitations associated with these applications. AI-generated content can contain inaccurate or fabricated information, misinterpret qualitative findings, reinforce existing biases, or overlook the nuances of participant experiences. Organizations are encouraged to establish clear safeguards and maintain meaningful human involvement throughout the evaluation process.
To support implementation, the resource includes a Recommended Practices checklist on page 6, organized around three stages of AI use: before, during, and after. It provides actionable guidance on staff training, selecting approved tools, protecting sensitive information, writing effective prompts, reviewing outputs, and documenting significant AI contributions. The guide also identifies practices to avoid, such as uploading confidential participant information into public AI platforms or using AI-generated findings without appropriate review.
This resource is particularly useful for:
- Nonprofits exploring how to incorporate AI into their evaluation and reporting practices.
- Evaluators and researchers seeking guidance on responsible AI use.
- Organizational leaders developing internal AI policies, procedures, or staff guidance.
- Teams working with sensitive participant information or community-based data.
- Program and communications staff using AI to organize information, develop materials, or communicate findings.
At BSRI, we believe technology should strengthen our ability to learn, collaborate, and make informed decisions while keeping people at the center of the process. When used thoughtfully, AI can help organizations improve efficiency, build evaluation capacity, and dedicate more time to meaningful work. Responsible adoption requires balancing these opportunities with a continued commitment to accuracy, equity, transparency, and the relationships that make evaluation valuable.
