OPM narrows high-impact AI rules for federal hiring
OPM is telling agencies that independent human review exempts hiring AI tools from pre-deployment testing and impact assessments, without defining what adequate review actually requires.
TL;DR
OPM issued guidance Thursday telling federal agencies that AI used in job announcements, applicant evaluation, file review, and hiring metrics generally isn't "high-impact" under federal AI governance rules, provided an independent human reviewer evaluates and adopts the output. The determination means agencies can deploy these tools without pre-deployment testing, impact assessments, or ongoing monitoring for adverse impacts. The memo doesn't define what constitutes sufficient independent review, and the Center for Democracy & Technology warns that tools shaping candidate assessments can meaningfully affect hiring decisions even when a human signs off.
OPM's Thursday memo to agency heads does more than clarify. It reverses the de facto position most agencies had taken on AI in hiring for two years: that nearly any AI touching employment decisions was "high-impact" under the OMB AI governance framework, triggering pre-deployment testing, impact assessments, and ongoing adverse-impact monitoring. OPM now says that was overcautious.
The memo's core move is its reading of "principal basis." Under the OMB framework, AI that serves "as a principal basis" for agency action is presumed high-impact, and federal employment determinations are explicitly listed among those presumptive categories. But OPM argues that when an agency official "independently reviews supporting data, evaluates the decisional logic, and adopts the AI-assisted work product, the output generally does not serve as a principal basis." The human reviewer doesn't merely override the designation. In OPM's framing, they "supply the decision's own justification."
The practical effect is substantial. AI tools that screen resumes, rank candidates, and generate interview recommendations can now be deployed without the minimum risk management practices that accompany a high-impact designation. No pre-deployment testing. No impact assessment. No requirement to monitor for adverse impacts.
What the memo doesn't address is what "independent review" actually demands. Must the reviewer understand the tool's decision logic? Are they expected to detect pattern-based disparate impact that an AI screening tool might embed? The memo doesn't say. Quinn Anex-Ries of the Center for Democracy & Technology flagged this as the central concern: a tool that summarizes candidates and issues recommendations "pretty logically will have a pretty big bearing on an employment decision," yet under OPM's framework that tool may still escape minimum safeguards.
OPM's language treats AI adoption as the default and governance as the friction to be reduced, a notable shift from the precautionary posture that characterized federal AI policy under the Biden-era executive order. CIO Adam Starr's accompanying blog post describes agencies as "overly cautious," and the memo itself warns that failure to adopt AI "may be compromising efficiency and quality" in federal hiring.
The memo's own reminder (that "a high-impact designation is not a prohibition on adoption") undercuts its premise. If the safeguards aren't a ban, the urgency to exempt hiring AI from them rests on a thinner justification than the memo provides.
Published ·Deep Fathom