OpenAI and its subsidiary Statsig have agreed to change how they recruit for jobs connected to employment-based immigration after a Justice Department investigation. The combined $3.2 million settlement addresses allegations that the companies steered opportunities away from U.S. workers while seeking to hire people with temporary employment visas.
On August 4, 2026, the Civil Rights Division announced the agreement. It covers fewer than 10 positions in the Permanent Labor Certification process, known as PERM, but imposes changes across job advertising, application handling, staff training and federal oversight. The agreement groups the San Francisco AI company with Statsig, its software subsidiary in Bellevue.
The announcement makes the distinction between an allegation and an operating requirement important here. The department described what its investigation found and the companies agreed to monetary and compliance terms. The agreement now sets concrete rules for future PERM recruitment, regardless of how few positions produced the case.
The case does not treat employment sponsorship itself as unlawful. PERM exists so companies can seek permanent residence for a foreign worker when good-faith recruitment does not find a qualified U.S. candidate. The allegation is that OpenAI and Statsig did not give domestic applicants a fair chance to enter that comparison. That makes the design and visibility of the recruitment process central to the settlement.
The Alleged Barriers Were Built Into the Application Route
PERM allows an employer to sponsor a worker for permanent resident status after making a good-faith effort to recruit qualified U.S. workers. The process is meant to establish that the sponsored hire is not displacing an available domestic applicant. Citizenship-status discrimination remains prohibited during that recruitment. Employers must test the labor market in good faith; the process cannot be structured to produce a preferred result by making the sponsored position harder for U.S. workers to discover or pursue.
According to the Justice Department, OpenAI did not place the affected PERM jobs on its external careers website even though it normally used that site for other openings. Applicants for the PERM roles were required to send paper applications by mail while candidates for other positions could apply electronically. The department also cited late-night radio advertisements among steps it said discouraged U.S. workers from applying.
Those details point to access, not merely the wording of a job description. A person cannot compete for a role they never see, and a paper-only route adds friction that does not exist for candidates using an online form. The settlement treats the recruitment channel itself as part of equal access to employment.
The $3.2 Million Is Split Between a Penalty and Back Pay
Of the total, $1.2 million goes to the federal government as a civil penalty. A separate $2 million fund is intended to provide back pay to people harmed by the practices alleged in the case. Combining those figures produces the headline settlement amount, but the two pools serve different purposes: one penalizes the violations alleged by the government, while the other is reserved for affected workers.
Engadget reported that the department had not explained publicly how it would identify recipients or calculate payments from the fund. That may be a difficult task because some potential applicants could have missed openings that were not displayed where they expected to find them. The department framed the remedy around access to highly sought-after technology jobs. The settlement announcement establishes the fund but does not provide a public list of eligible workers. It also leaves open practical questions about outreach to people who did not know the roles existed, the evidence they would need and the timing of any distribution.
The number of positions does not translate directly into the number of possible recipients. No per-person amount has been published. The department said the amount was meant to reflect lost access to well-paid roles in the technology industry. The $2 million fund is therefore a compensation pool, not a published payment schedule.
OpenAI Must Put Future PERM Jobs on Its Public Careers Site
The practical changes begin with visibility and format. For later PERM recruitment, OpenAI's own careers page must carry the openings, and candidates must be able to submit online. It must also revise employment policies, train relevant personnel on the Immigration and Nationality Act's anti-discrimination requirements, and submit to Justice Department monitoring and reporting. These are prospective duties: they govern how the company handles later recruitment instead of ending with the transfer of settlement money.
Those obligations turn a dispute over a small set of jobs into a continuing compliance program. Public listings and electronic submissions create records that can be compared with the company's ordinary hiring process. Training addresses the people running recruitment, while monitoring gives the department a way to assess whether the agreed changes last beyond the settlement date.
The case is the department's 13th settlement since it relaunched the Protecting U.S. Workers Initiative in 2025. The department says agreements under that program can combine civil penalties, back pay, recruiter training and an end to unjustified citizenship-status restrictions. For OpenAI, the lasting consequence is not just the $3.2 million outlay. Its immigration-linked recruiting must now use the same visible, accessible pathways that U.S. applicants would reasonably expect from a major technology employer.