The AAO dismissed an EB-1A appeal by an AI research scientist who met four evidentiary criteria but could not show…

A research scientist specializing in machine learning, computer vision, and generative AI sought EB-1A classification. SCOPS found four of the ten criteria satisfied (judging, original contributions, scholarly articles, and high salary) and proceeded to a final merits determination, ultimately denying the petition. On appeal, the AAO agreed that the petitioner's award, press coverage, peer review activity, citation metrics, salary, and recommendation letters — while reflecting some accomplishment — did not collectively demonstrate sustained national or international acclaim or placement among the small percentage at the very top of the field. The AAO emphasized that contextualizing evidence against peers at the top of the field is essential, and that general letters, nominal press mentions, and citation data without self-citation filtering are insufficient to clear that bar.

What worked: The petitioner cleared the initial evidentiary threshold by satisfying four criteria — peer review service, original contributions via citation metrics, scholarly authorship, and documented salary — allowing the case to proceed to a final merits review.

What failed: Evidence failed at the final merits stage on multiple fronts: (1) the fellowship award lacked demonstrated prestige or field-wide recognition; (2) press coverage was nominal and did not mention the petitioner by name; (3) peer review activity was not distinguished from ordinary participation or compared to top researchers; (4) citation metrics lacked self-citation filtering and context showing top-of-field placement; (5) recommendation letters were too general and did not compare the petitioner to others in the field.

Takeaway: Researchers in competitive fields like AI must go beyond simply meeting the initial criteria — they must quantitatively and qualitatively compare themselves to peers at the very top of the field using benchmarks, rankings, and expert letters that directly address relative standing. Evidence submitted without that comparative context will be treated as insufficient at the final merits stage regardless of raw numbers.

Cases like this are frequently used by attorneys when responding to RFEs or building initial petitions. The evidence patterns that worked (or failed) here directly reflect what USCIS officers look for when evaluating EB-1A criteria.

Criterion-by-criterion breakdown