On 25 February 2026, the US Department of Justice announced a settlement with Elegant Enterprise-Wide Solutions. According to the announcement, the company published AI generated job advertisements with citizenship status restrictions not authorized by law, including language limiting consideration to H-1B, OPT, or H-4 categories.
This article summarizes a public announcement for business learning and is not legal advice.
The cost goes beyond a penalty
The announcement identifies a $9,460 civil penalty and compliance terms including training and monitoring over three years. Updating processes and maintaining oversight also require organizational time.
Match the control to the publication
Before a job advertisement is published, an authorized reviewer should check eligibility language against appropriate legal guidance. Use approved wording for sensitive terms rather than asking a model to invent them.
Retain the approved text and reviewer identity. A suitable record of instructions and tool version, consistent with retention policy, can help explain how an error occurred and how to correct the process.
Working across borders
Recruitment rules differ between jurisdictions. Teams sharing tools and templates should confirm that the language fits the role and location. A common review process can help, but one template should not be assumed appropriate everywhere.
Can AI still support drafting?
Drafting can be considered with suitable review and approval. Our takeaway from this matter is that responsibility for publication must remain clear when the first draft is generated automatically.