Donate
Independent, objective, nonpartisan research

PPIC Guidelines for Use of Artificial Intelligence

July 22, 2026

Generative artificial intelligence (AI) can be defined as “computational techniques that are capable of generating seemingly new, meaningful content such as text, images, or audio from training data” (Feuerriegel et al., 2024). Generative AI differs from other types of AI because it can produce its own original, humanlike output rather than merely recognizing patterns in data to perform a specific task (Marr, 2023).

PPIC is a nonprofit, nonpartisan think tank with a mission to inform and improve public policy in California through independent, objective, nonpartisan research. Our primary responsibility is to maintain the independence and rigor of our peer-reviewed work. Our approach to AI seeks to ensure that nothing compromises that commitment to quality and transparency.

We recognize that generative AI has the potential to improve the efficiency and productivity of PPIC workflows, but we do not require or expect staff to use AI. Our teams use the tools for coding, drafting outlines and basic text, summarizing written content, conducting literature reviews, and gathering information. We do not use AI tools with sensitive data, and all content produced with AI is reviewed and validated.

Our AI policy applies to the following requirements:

  • Information privacy and security. Staff are prohibited from uploading any personal identifying information or any sensitive or restricted data to a generative AI. All staff or consultants who use generative AI for PPIC work must turn off the data sharing function, so that their prompts are not used to train future models.
  • Content review. Staff must carefully review any content generated by AI that is used in externally facing work. This includes independently verifying all citations and facts. If staff use AI for a task in which each piece of information is not independently verified—such as data collection—the user must establish a methodology to minimize the error.
  • Disclosure. For content that will go through normal human review, staff are not required to disclose AI use externally; staff are encouraged to share use internally to improve processes and understanding. However, in cases where it is difficult or impossible to independently verify AI-generated content—for example, where the AI has scraped a large amount of information from the web—the use of AI and the methodology for minimizing error must be disclosed in the final publication.
  • Uphold PPIC’s values. Employees at PPIC must always be mindful of PPIC’s reputation and values as they do their work, and work with AI is no exception. PPIC’s values include a commitment to environmental leadership and an appreciation for California’s diversity, so employees are encouraged to consider the broader impacts of AI on these goals when considering using the tools.