The Bridgespan Group publishes research reports, articles, and other resources with the goal of helping social sector leaders and funders strengthen their organizations and practices as well as to encourage more funding to flow toward equitable social change. 

We have found that generative AI tools can be helpful in our publishing work, for example, in synthesizing and organizing large amounts of research. Facts and accurate data are the bedrock of Bridgespan’s insights and recommendations, and the trust-based relationships we have with the social sector. So we take steps to mitigate and manage risks associated with generative AI. Ultimately, Bridgespan is responsible for the accuracy of all published content.

We do not use generative AI to draft or substantively edit our reports, articles, and other resources. From time to time, we may use generative AI to produce derivative content (e.g., social media posts, newsletter blurbs, press releases) from our reports and articles, only drawing on information from our originally published research.

We may have used generative AI to assist with research, including identifying source material, synthesizing primary and/or secondary research, analyzing data, and providing initial outlines for consideration. We apply the following principles in all instances:

• We use only generative AI tools that we have formally vetted for data protection and security.
• We only use enterprise-grade tools that keep data secure and that do not use our data to train models. 
• We test for data quality and bias by having Bridgespan employees:
  1. Treat every output as a draft that needs revision
  2. Assume that data generated are incorrect, especially quantitative outputs or recent events, and validating data through more robust research
  3. Actively employ strategies to mitigate bias, especially around populations, geographies, or perspectives underrepresented in datasets
  4. Take responsibility for the accuracy, quality, and integrity of any deliverable that incorporates AI assistance
• We do not use generative AI tools to create or modify video or photography. 
• We will disclose all use of generative AI that deviates from these principles.