DOJ Slams OpenAI—$3.2M Over Hiring Trick

The Justice Department’s $3.2 million settlement with OpenAI over how it filled a handful of sponsored positions is a clear signal: under the Trump administration, immigration-linked hiring practices that quietly sideline U.S. workers are now squarely a civil-rights enforcement priority, especially in the tech sector.

Key Points

OpenAI and its subsidiary Statsig agreed to pay $3.2 million after DOJ alleged they favored temporary visa holders over U.S. workers in PERM-sponsored roles.
DOJ’s investigation found those PERM jobs were handled differently from other openings, including not being posted on OpenAI’s public careers site and using application methods that discouraged U.S. applicants.
The settlement requires OpenAI to overhaul recruitment: public posting of PERM roles, acceptance of electronic applications, staff training, and ongoing DOJ monitoring.
This case sits within a broader Trump-era push to police employer use of foreign-worker visa programs and AI-related hiring tools to protect U.S. workers.
The enforcement message is straightforward: if you rely on immigration programs or automation in hiring, you remain fully accountable under federal anti-discrimination law.

What DOJ Found in OpenAI’s PERM Recruitment

The core of the government’s case against OpenAI is not that it used the permanent labor certification process—known as PERM—but how it used it. PERM itself is a lawful mechanism that allows employers to sponsor foreign workers for permanent residence when they can show there are no available, qualified U.S. workers for the job. The Civil Rights Division’s allegation is that OpenAI and Statsig structured their recruitment so that U.S. workers never had a fair shot at those sponsored roles.

According to DOJ, the companies routinely advertised ordinary roles on their public careers website and accepted standard online applications. For PERM-linked positions, however, they allegedly departed from that norm. Axios and DOJ materials describe several practices: not posting PERM positions on the external job site at all, requiring paper applications sent by mail rather than the usual online submissions, and advertising some openings in channels such as late-night radio that are unlikely to reach the typical pool of highly skilled U.S. tech workers. In effect, the agency concluded, these choices made it less likely that qualified Americans would see, much less apply for, the jobs that were already earmarked for foreign workers on temporary visas.

Assistant Attorney General Harmeet K. Dhillon summarized the department’s position bluntly: “It is illegal to discriminate against U.S. workers by preferring temporary visa holders for jobs,” and the settlement is designed to “ensure that OpenAI redresses harm and changes its recruitment practices so that U.S. workers receive a fair opportunity for highly sought-after technology positions.”

Financial Terms and Structural Remedies

The settlement’s dollar figure—$3.2 million—looks large next to the small number of roles involved. DOJ and subsequent reporting indicate that fewer than ten positions were at issue, yet the resolution includes $1.2 million in civil penalties and $2 million reserved for a victim fund to compensate U.S. workers who lost opportunities. That ratio is deliberate. Civil-rights settlements in employment are often structured so that the financial impact reflects not just the discrete headcount but the principle at stake: shutting U.S. workers out of gateway jobs in a cutting-edge field is treated as serious economic harm, even when the numbers are small.

Beyond money, the decree imposes concrete operational changes. OpenAI must now post PERM positions on its careers site in the same way it posts other roles, accept electronic applications instead of forcing mail-in paper submissions, train relevant recruiting and HR staff on anti-discrimination requirements, and submit to monitoring and reporting obligations so DOJ can verify that practices actually change. These requirements are typical of modern consent settlements: they aim to reshape systems, not just penalize past behavior.

It is also significant that OpenAI chose to settle rather than litigate to a merits ruling. The Wall Street Journal notes that while the company disagreed with DOJ’s findings, it resolved the matter through settlement. That posture is common in PERM enforcement: companies often prefer compliance commitments over years of contested litigation, particularly when reputational stakes are high.

How PERM Cases Became a Civil-Rights Priority

To understand why this case drew such a strong response, you have to situate it in a larger enforcement pattern. Over the past several years, the Civil Rights Division has relaunched and expanded an initiative focused on protecting U.S. workers from misuse of foreign-worker visa programs, including PERM and various temporary visas. The Journal notes OpenAI is the most prominent firm to settle since this renewed push began.

These matters share a structural theme: DOJ does not challenge the existence of PERM sponsorship itself; instead, it looks for ways employers treat PERM jobs differently from other openings. When PERM roles are advertised in obscure channels, require unusually burdensome application methods, or are essentially invisible to ordinary U.S. job seekers, the department reads that as evidence that the process is being used to lock in a specific foreign candidate while technically complying with labor-certification paperwork. Recent commentary from employment lawyers places the OpenAI settlement alongside earlier cases involving large tech companies such as Apple and Meta, suggesting an emerging playbook: a pattern of differential treatment for sponsored roles, followed by a negotiated settlement that combines penalties, victim funds, and practice changes.

This enforcement direction meshes with the broader Trump administration emphasis on curbing what it casts as abuse of immigration programs by large employers. Public signals—from speeches to policy guidance—have stressed that high-skilled visa pathways must not become back doors for sidelining domestic talent in favor of lower-visibility sponsored workers. In that context, a high-profile AI company becomes both a symbolic and practical test case.

OpenAI’s Position and the Limits of Settlement

The DOJ announcement and associated reporting make clear that the department views this as a discriminatory pattern, but settlement documents stop short of a judicial finding of liability. The Wall Street Journal reports that OpenAI disagreed with the DOJ’s conclusions, yet opted to settle to resolve the matter. That is standard in civil-rights employment enforcement: consent agreements generally include no admission of wrongdoing, even when they carry meaningful penalties and oversight.

For observers, the important distinction is between legal posture and operational reality. OpenAI can maintain that its intent was not to disadvantage U.S. workers, but the company has nonetheless committed to change the practices DOJ identified as problematic—changing how PERM jobs are posted, which application methods are allowed, and how staff are trained and monitored. From a compliance perspective, what matters is the new baseline: PERM roles must now be visible and accessible to U.S. applicants on equal terms.

AI, Hiring Bias, and Legal Accountability

This OpenAI case is formally about immigration-linked recruitment, not algorithmic decision-making, but it lands in a moment when AI and hiring bias are already under scrutiny. Academic work highlighted in HR and legal circles shows that large language models used in simulated hiring scenarios can independently develop novel social biases when they are rewarded only for predictive success, leading to occupational segregation far beyond human stereotyping. Separate studies have found that résumé screeners built on GPT-type models can favor white-associated names, underrate women’s experience, and misattribute immigrant status to Asian and Hispanic candidates, with bias patterns varying by model, occupation, and configuration.

Federal agencies have responded by reiterating that longstanding civil-rights statutes—Title VII of the Civil Rights Act, the Americans with Disabilities Act, the Age Discrimination in Employment Act—apply fully to AI-enabled hiring tools. The EEOC has brought guidance and enforcement actions against employers deploying automated screens that produce disparate outcomes; the ongoing Mobley v. Workday litigation underscores that vendors cannot simply declare themselves “decision-support” and escape scrutiny when their tools drive systematic rejections of protected groups.

Seen against that backdrop, a settlement focused on PERM hiring sends a complementary message: whether the mechanism is a visa program or an algorithm, employers are responsible for the structures they build around it. If those structures make it harder for certain categories of U.S. workers to see jobs, apply for them, or be fairly assessed, DOJ and other agencies are prepared to intervene.

OpenAI just cut DOJ a $3.2M check over hiring bias — for fewer than 10 jobs. $1.2M penalty, $2M victim fund. DOJ says OpenAI ran midnight radio ads and paper-only apps to steer Americans away from visa-favored roles. 13th such DOJ settlement this year.

— WhodeySPOH (@WhodeyAI) August 4, 2026

What This Means for Employers Using Immigration Programs and AI

For employers—especially in technology and other high-skill sectors—the OpenAI settlement is best read as a practical compliance roadmap. First, PERM and other sponsorship-related recruitment must be integrated into ordinary hiring channels. If a company typically posts roles on a public careers site and accepts online applications, sponsored positions should follow the same pattern; carving out opaque or burdensome processes for those jobs invites scrutiny.

Second, optimize for fairness, not just throughput. The research on AI hiring tools shows how easily models can develop or amplify bias when they are tuned purely for efficiency or predictive success. Employers need clear internal documentation of what each automated system is optimizing for, continuous testing of outcomes across protected groups, and meaningful human oversight with authority to challenge or override AI recommendations. Immigration-linked hiring requires a similar discipline: recruiters and lawyers must coordinate to ensure that compliance steps do not become de facto barriers for U.S. candidates.

Third, transparency matters. Both DOJ’s settlement approach and the emerging case law on AI hiring emphasize notice and recourse. Candidates should know when automation or sponsorship structures are involved in their assessment, and they should have a path to request reconsideration or human review. In the PERM context, that translates into clear public information about sponsored roles and straightforward application methods that do not privilege insiders.

Finally, the OpenAI matter shows that scale is not a shield. Fewer than ten roles triggered a multimillion-dollar settlement and national attention. For executives, the lesson is simple: small pockets of noncompliant practice can carry outsized risk when they intersect with visible policy priorities—protecting U.S. workers, policing hiring bias, and regulating AI. Building robust, fair systems from the outset is far less costly than redesigning them under the eye of federal enforcement.