Labor & Employment Law Daily Wrap Up, PUBLIC EMPLOYEES—Trump administration will employ AI to review, analyze personnel records of federal employees, (Apr 10, 2025)
Organizations Mentioned:Office of Personnel Management
By Patricia K. Ruiz, J.D.
The AI will employ stateless models to ensure that inputs and outputs are not used to train, retrain, or improve base models.
The Office of Personnel Management (OPM) announced its plans to utilize artificial intelligence (AI) to extract various types of data from forms contained in the federal government’s electronic official personnel folder system (eOPF), for the purposes of analyzing and verifying the accuracy of data and updating inaccurate information, according to a privacy impact analysis (PIA) released by OPM. The PIA gives an overview of the actions being taken to protect employees’ information within the eOPF, both generally and for purposes of the AI use.
The AI project reflects an executive order from the Trump administration revoking “certain existing AI policies and directives that act as barriers to American AI innovation,” to allow the U.S. to retain global leadership in AI.
eOPF. Federal employees each have a single personnel folder, known as the OPF, which documents their entire federal civilian service. The eOPF contains digital images of federal employees’ paper OPFs, virtual work folders for use by human resources (HR) offices, and folders for non-federal employees that are accessible online. According to the OPM, the OPF contains records needed by the government to make accurate employment decisions throughout an individual’s federal career. The records can be used to show that a federal appointment was valid or verify military service credit for leave, reduction-in-force, or retirement. The records contain an individual’s employment history, grades, occupations, and pay and contain document choices made by the employee under federal benefits programs.
The eOPF includes an audit trail that records when and why an individual reviewed a folder within the eOPF. The audit logs are periodically reviewed to ensure information is being access and used appropriately.
Access. The eOPF is available to authorized users via the internet; employees may access the eOPF to view their own OPF but cannot modify their files or view other employees’ OPFs. HR specialists may only access the OPF and virtual work folders of employees to whom they provide HR services, and Government officials who need to see the records in the eOPF to do their jobs also have access. The eOPF employs role-based access controls and permits access to employees according to their role. The agencies that submit information to the eOPF determine which of their HR employees require access. All OPM employees and contractors who support the eOPF are required to complete OPM-provided, mandatory, annual security and privacy awareness training prior to gaining access to the eOPF.
AI use. OPM explained that any use of AI will be done within a controlled environment, utilizing systems that have attained FedRAMP High authorization and will be performed by approved OPM officials or qualified federal government contractors. The AI will employ stateless models to ensure that inputs and outputs are not used to train, retrain, or improve base models. Before AI is deployed in the system, comprehensive audit checks will be conducted to verify the accuracy of any work performed.
Notice, consent. Federal employees will receive prompt notice by email of any changes to their records in the eOPF that result in the issuance of a corrected form. Individuals may decline to provide information on the various forms entered into the eOPF, though there may be consequences for not providing certain information. Once individuals provide information on those forms, the federal agencies using the eOPF determine what information is sent to the eOPF. Employees have no ability to consent to particular uses of their information in the eOPF, and there is no opportunity for individuals to decline to have their information included in the system.
News: AgencyNews IndustryNewsTrends PublicEmployees Privacy AINews