AI for policy
ESN-CA is committed to incorporating AI safely and responsibly in its contributions to the evidence-support system, which we often shorten to ‘AI for policy.’
We are doing so in several ways:
- incorporating AI safely and responsibly in the existing evidence-synthesis workflows used by our evidence-synthesis teams, such as in the rank-ordering of studies being considered for inclusion in an evidence synthesis
- gearing up to incorporate AI safely and responsibly in the transformed workflows that the Evidence Synthesis Infrastructure Collaborative (ESIC) is now designing, including the movement to continuous evidence surveillance (daily ingestion of newly published empirical studies from OpenAlex and mapping to sector and then to a sector-specific taxonomy), ‘living evidence maps,’ and policy-scale living evidence syntheses
- contributing to ESIC’s efforts to shift the trajectory of ‘AI for policy’ large-language model-derived tools (e.g., evidence synthesis, localization assessments, and mixing and matching needed forms of evidence)
- piloting the use of an AI bot that leverages the Progress Agentic RAG (retrieval-augmented generation) and a ‘walled garden’ of our and other trusted groups’ resources to respond to queries.