State Department Testing Machine Learning Model to Improve FOIA Processing

Eric Stein, U.S. Department of State deputy assistant secretary for the Office of Global Information Services, said the department is looking for ways to quickly process Freedom of Information Act requests using automation tools and artificial intelligence.

The federal government received over 900,000 new FOIA requests in fiscal year 2022, prompting the Chief FOIA Officers Council’s Technology Committee, which Stein co-chairs, to look for ways to better understand the tools being used by agencies to process the requests.

Stein said the processing of FOIA is still mostly being done manually, Federal News Network reported.

The State Department’s annual FOIA report for FY 2022 indicated 14,000 new FOIA requests and another 21,000 pending requests.

Stein said the use of a trained machine learning model to improve the declassifying of records is also being piloted by the department. The ML model is comparable to the accuracy of human professional FOIA processors 97 to 99 percent of the time and has saved the department work equivalent to half a year.