<p>Recent U.S. government electronic recordkeeping policies and practices have resulted in a vast accumulation of preserved White House and federal agency email records in the billions of pages, all of which are subject to public access under the Freedom of Information Act (FOIA). Notwithstanding the enormity of the task, at present government personnel continue to rely on a combination of keyword searching and manual review to find responsive records and redact material that agencies may withhold under one or more FOIA exemptions from disclosure. In confronting similar demands for access to electronically stored information, the legal profession has adopted AI in the form of technology assisted review (TAR) methods to greatly expedite searches for responsive records to e-discovery requests. Adoption of TAR search methods by access professionals would help eliminate backlogs and improve government transparency. Moreover, recent research applying text classification methods to identify exempt material have shown a measure of success in one use case involving the segregation of documents and portions thereof containing material covered under the “deliberative process privilege.” In a further limited extension of the research, generative AI in the form of ChatGPT was used to construct narratives explaining the reasoning for the decision to withhold deliberative process materials, in a manner that supports future use of generative AI as an intelligent assistant aiding in human review.</p>