The US Justice Department announced on 4 August 2026 that OpenAI and its subsidiary Statsig will pay $3.2 million to settle claims they favored foreign workers on temporary employment visas and discriminated against US job applicants in recruiting and hiring (The Guardian).

The settlement breaks down into $1.2 million in civil penalties and $2 million in compensation for alleged victims. Both companies were accused of violating the Immigration and Nationality Act, the federal law that bars employment discrimination based on citizenship status.

At the center of the case were fewer than 10 positions, according to the Justice Department. The DOJ's settlement agreement identifies a "pattern or practice of citizenship status discrimination in five PERM-related job advertisements." PERM stands for Program Electronic Review Management — the process through which employers sponsor foreign workers for permanent residency (a green card). Before a company can sponsor someone, PERM requires the employer to show it tested the US labor market and found no qualified, willing, and available American workers for the role (DOJ Settlement Agreement).

The alleged tactics were deliberate and specific. The Justice Department said OpenAI and Statsig recruited foreign workers for certain open positions while taking steps to discourage US applicants: requiring paper applications instead of electronic ones, advertising jobs on late-night radio, and declining to post openings on any external website (The Guardian).

OpenAI denied wrongdoing in the settlement agreement. Despite the denial, the company agreed to revise its employment policies, conduct training, and submit to DOJ monitoring of its recruitment practices.

Assistant Attorney General Harmeet Dhillon framed the settlement as both corrective and deterrent, saying it ensures OpenAI "redresses harm and changes its recruitment practices so US workers receive a fair opportunity for technology positions" (The Guardian).

The case does not exist in isolation. The Justice Department has announced at least a dozen other settlements over the past year involving alleged discrimination against US workers, predominantly by technology companies. None of those prior settlements carried the profile of the OpenAI matter, which links the most prominent name in artificial intelligence to the politically charged debate over temporary work visas.

President Donald Trump has repeatedly asserted that companies abuse the temporary employment visa system and has pushed to restrict the hiring of foreign workers. His administration imposed a $100,000 fee on new H-1B visas for highly skilled workers, though that measure has been blocked pending legal challenges. The OpenAI settlement fits squarely within that enforcement and policy posture, even though the case was brought under a citizenship-discrimination statute rather than visa-reform authority.

The PERM process referenced in the DOJ's settlement agreement is the procedural gateway employers use to demonstrate they have tested the US labor market and found no qualified, willing, and available American workers before sponsoring a foreign national for permanent residency. When the DOJ alleges discrimination "in PERM-related job advertisements," it is targeting the very recruitment steps that are supposed to give US workers a fair shot — the job postings, application methods, and advertising channels that constitute the labor market test.

The relatively small number of positions at issue, fewer than 10, stands in tension with the $3.2 million settlement figure. The Justice Department explicitly addressed that disparity, stating that the sizeable settlement reflected the harm OpenAI's conduct caused to US workers. This signals the DOJ's willingness to calibrate penalties to company profile and conduct severity rather than to a strict arithmetic of affected positions.

For technology employers, the settlement underscores the DOJ's enforcement focus on recruitment-stage discrimination, not just hiring outcomes. The specific tactics cited — analog application methods, restricted advertising channels, suppressed job postings — describe a methodology for funneling candidates rather than overt exclusion. Companies that maintain PERM programs will need to scrutinize whether their recruitment advertising practices could be construed as engineered to deter any applicant pool.

The broader context here is a convergence of enforcement priorities and political pressure on the temporary visa system. With the H-1B fee blocked in court, the DOJ's citizenship-discrimination docket may be the more durable mechanism for the administration's policy goals. The OpenAI settlement, combining a high-profile target with structural remedies including monitoring and training, sets a template that could inform the department's approach to the remaining tech-sector cases in its pipeline.