Overview

This chapter explains how the OpenAI–Statsig settlement makes PERM recruitment observable after an application enters the system. It requires good-faith, lawful, job-related review; permits comparable interviews or assessments when used for comparable non-PERM positions; and adds semi-annual position-level reports, training for people involved in PERM recruitment, policy review, IER inquiries, and notice-and-cure. For reporting, qualified means meeting the applicable PERM standard, not automatically the employer’s ordinary hiring standard. These controls create evidence and correction paths, but the sources do not show completed compliance, applicant-level decisions, or hiring outcomes.

Learning outcomes

Applicants may be rejected only for lawful, job-related reasons, with comparable assessments used consistently for comparable non-PERM positions.
Semi-annual counts connect positions to recorded applications and PERM-qualified applicants, but they are not self-interpreting fairness scores.
Training, policy review, IER inquiries, and notice-and-cure make compliance more observable; the record does not show later results.

Keywords

OpenAIStatsigPERMReviewTrainingOversight

Making Fairness Observable: Review, Training, and Oversight

The OpenAI and Statsig settlement treats an equal application route as the beginning of fair PERM recruitment, not the end. Once a U.S. worker can reach and submit an application, the organization still has to review that person under lawful, job-related standards, keep the application visible, train the people who operate the process, and produce enough information for oversight. This chapter focuses on that operating loop and on its limits: the agreement creates evidence and correction mechanisms, but it does not guarantee a hiring outcome or prove that later compliance has already occurred.

Review applicants on job-related and comparable terms

The settlement requires the Respondents’ PERM policy to direct a good-faith review of the qualifications of applicants for future PERM-related positions. It says U.S. workers may be rejected only for lawful, job-related reasons and requires that Respondents not discriminate against any applicant based on citizenship status when complying with 20 C.F.R. § 656.10(c). These provisions answer a question that an equal-access policy alone cannot answer: what happens after an applicant enters the system?

A fair route into consideration does not mean that every applicant is qualified for the job or that every applicant receives the same final decision. It means the reason for a rejection must be connected to the job and lawful, rather than to citizenship status. The settlement’s wording also requires good-faith review, which directs attention to whether the application was genuinely considered rather than merely collected to satisfy a formality. The source does not provide a test for proving good faith in an individual case, so the article should not pretend that the phrase supplies an automatic result. It does establish the standard the Respondents agreed to put into their PERM policy.

The agreement permits job-related inquiries, including technical interviews and assessments such as testing coding ability, when the same interviews or assessments are used for comparable non-PERM positions. That condition is a comparator safeguard. A technical assessment may be a legitimate way to evaluate whether an applicant can perform the required duties at the level required for a position. But the settlement says it should be used consistently, based on the requirements of the PERM-related position, and only where the same kind of interview or assessment is used for comparable non-PERM positions.

The comparison does important work. It prevents a company from treating the PERM label as a reason to invent a special screening burden for U.S. workers. At the same time, it does not require unrelated jobs to have identical assessments. The relevant comparison is between a PERM-related position and a comparable non-PERM position, using a job-related tool for the duties and level required. The sources do not identify the companies’ actual tests, the positions to which they were applied, or whether a particular applicant passed or failed. They support a rule about the basis and consistency of evaluation, not a conclusion about any individual’s qualifications.

The settlement gives “qualified” a specific meaning for one reporting purpose. In the semi-annual report, the term relates to an application submitted by a U.S. worker who met the applicable PERM qualification standard under 20 C.F.R. § 656.17(g)(2). It must not be interpreted to mean that the individual met the Respondent’s ordinary hiring standard. This distinction prevents two opposite errors.

Published by DeepMox

FAQ

What does good-faith review require under the settlement?

It requires review of applicants’ qualifications and permits rejecting U.S. workers only for lawful, job-related reasons; it does not provide a test that proves good faith in every case.

What does qualified mean in the settlement’s semi-annual reports?

For reporting, qualified means that a U.S. worker met the applicable PERM standard under 20 C.F.R. § 656.17(g)(2), not necessarily the Respondent’s ordinary hiring standard.

Can application counts prove that a PERM process was fair?

No. Counts record applications and qualification totals for positions, but they do not show people who never entered the system, equal visibility, the reason for a low count, or the absence of discrimination.

Previous10 minOpenAI–Statsig PERM Recruitment Controls

Next13 minOpenAI–Statsig PERM Compliance Chain

OpenAI–Statsig PERM Review and Oversight

Making Fairness Observable: Review, Training, and Oversight

Review applicants on job-related and comparable terms

FAQ