The Office of Personnel Management (OPM) is advising federal agencies to expand the use of artificial intelligence (AI) throughout the hiring process while applying risk controls based on how the technology is used in employment decisions.
In an Aug. 27 memorandum, OPM said agencies should use AI to help advance governmentwide goals for better position descriptions and assessments, greater hiring efficiency, and reducing average time-to-hire to 80 days.
The guidance applies the Office of Management and Budget (OMB) Memorandum M-25-21 high-impact standard to common hiring use cases. Under that policy, an AI use can be high impact when its output serves as a principal basis for a decision with a significant effect on access to employment – including “[d]etermination of the terms or conditions of Federal employment, including pre-employment screening, reasonable accommodation, pay or promotion, performance management, hiring or termination, or recommending disciplinary action; reassignment of workers to new tasks or teams.”
OPM said human review alone does not automatically make an AI use lower risk. Instead, when an appropriate agency official independently reviews supporting data, evaluates the decisional logic, and adopts the AI-assisted work product, OPM generally views the AI output as not serving as a principal basis for the resulting decision or action.
The memo identifies four categories of AI use that OPM generally considers not high impact under M-25-21: developing content for job announcements, supporting applicant-level evaluations, conducting pre-offer quality control, and analyzing aggregate hiring programs.
However, agencies must assess their specific workflows and deployment contexts to reach a non-high-impact determination, and an appropriate agency official must document that determination to the agency’s chief artificial intelligence officer (CAIO).
Content-development uses may include drafting position descriptions, job analyses, qualification requirements, interview questions, and other hiring materials. Before operational use, agencies must review, validate, and adopt those materials under applicable personnel requirements.
For résumé screening, eligibility and qualification reviews, and candidate summaries, OPM said AI may locate, extract, and summarize information from applicant documentation without serving as the principal basis for a decision if the deciding official independently reviews the underlying record and makes the determination on that basis.
If officials instead ratify AI outputs without independently reviewing the underlying record, OPM said the AI output serves as a principal basis for the decision and the use must be treated as high impact.
Regardless of whether an AI use is considered high impact, agencies must verify that AI-assisted rating, ranking, and referral correctly apply veterans’ preference; that AI-enabled steps meet accessibility and reasonable-accommodation requirements; that applicant data is processed only in agency-authorized tools with applicable privacy protections; that hiring actions remain documented and auditable; and that existing reconsideration avenues remain available for adverse eligibility determinations.
For AI uses that remain high impact, agencies must implement OMB’s minimum risk-management practices or submit a waiver request to their CAIO for one or more of those requirements.
OPM also stressed that a high-impact designation does not prohibit agencies from adopting an AI use. The agency encouraged chief human capital officers to work with CAIOs and other relevant stakeholders, including chief privacy officers, chief information officers, and chief data officers, to implement the applicable safeguards.
OPM said it will continue introducing AI features across its USA Suite of tools, including USA Staffing, USAJOBS, USA Hire, USA Class, and USA Performance. However, the agency said agencies should not wait for those features before adopting AI consistent with the memorandum.
OPM also plans to address AI-assisted processes through its delegated examining oversight and provide assistance to agencies through its Artificial Intelligence Governance and Adoption Program.