Legal Issues with the Use of Artificial Intelligence to Fulfill FOIA Requests

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Several federal agencies are piloting the use of artificial intelligence (AI) to fulfill federal Freedom of Information Act (FOIA) requests, both in finding documents and redacting non-disclosable information. As agencies develop AI systems for this purpose, they must also meet the established legal requirements for searches and responses. This paper first examines the legal requirements for processing FOIA requests as determined by the statute, implementing regulations and guidance, and court rulings. It then examines the nascent processes that are being piloted and/or proposed to use AI for FOIA processing and how these processes comport with the legal requirements. Finally, we propose protocols to ensure that the legal requirements of FOIA are met in the use of AI to fulfill requests under the law.