AI and ML research can produce strong evidence for EB-1A — top-venue conference authorship, peer-review service, citation impact, widely-adopted models — but how the petition is assembled matters as much as the underlying record, and outcomes are case-specific.
This page covers how AI and machine-learning researchers pursue the EB-1A extraordinary-ability green card, a self-petition needing no employer sponsor. Qualification requires either a major international award or at least three of ten criteria under 8 C.F.R. § 204.5(h)(3); top-venue authorship, peer review, and citation impact are the criteria that typically carry these records.
Most AI-researcher EB-1A petitions we prepare come from senior individual contributors, staff or principal research scientists at industry labs, tenure-track or tenured faculty, and postdoctoral researchers with a strong publication and citation record. Subfields skew toward natural language processing, computer vision, generative AI, reinforcement learning, robotics, and ML systems — anywhere there is a recognized top-tier conference circuit.
Some applicants come to us expecting that their CV alone will carry the petition. Others come worried that their work is too applied, too industry-flavored, or too recent to register as extraordinary. Both groups benefit from the same up-front exercise: a candid look at how each piece of evidence might map to a USCIS criterion, what is missing, and where the case is genuinely strong before a draft is written.
If you are early in your career, still in graduate school, or your first papers are in lower-tier workshops, EB-1A may be premature. We will tell you that on the strategy call rather than at the end of a six-month engagement.
Best Paper and Outstanding Paper Awards at NeurIPS, ICML, ICLR, ACL, EMNLP, CVPR, ICCV, AAAI, and comparable top-tier venues tend to carry the most weight as awards evidence in this field. Honorable mentions and Best Paper Runner-Up recognitions can also support the criterion, particularly when the conference is itself competitive. Workshop-level awards are typically weaker on their own and usually need to be paired with other evidence. We document the selection process — submission count, acceptance rate, and selection committee composition — because USCIS frequently questions whether an award is nationally or internationally recognized. Whether any particular award satisfies the criterion is decided case-by-case by the adjudicating officer.
IEEE Senior Member status is among the more common forms of membership evidence we see for AI researchers, and Distinguished or Fellow grades at IEEE or ACM tend to be stronger. Standard ACM or IEEE membership generally has not been treated as satisfying this criterion because there is no outstanding-achievement requirement for entry. Invited membership in selective programs — for example, a long-term institute affiliation that requires nomination and committee review — has supported this criterion in cases where the entry standards can be documented. Whether a given membership qualifies depends on the association's published criteria and how those criteria are characterized to USCIS.
AI researchers tend to underweight this criterion because their work is published by them, not about them. Evidence that has supported this criterion in past cases includes media coverage of the researcher or their work — for example, a TechCrunch or Wired profile of a model release, a New York Times piece quoting the researcher on a policy issue, or an interview on a major podcast about a specific technical contribution. Conference talk recordings are generally not treated as third-party coverage. Personal blogs and the researcher's own posts typically do not qualify. The strength of any particular media item depends on the publication, its reach, and the substance of the coverage.
Reviewing for top venues — NeurIPS, ICML, ICLR, ACL, EMNLP, CVPR, ICCV, AAAI, KDD, and comparable conferences — has supported this criterion in many cases, as has service as Area Chair, Senior Program Committee Member, or Workshop Organizer. Journal review for venues such as JMLR, IEEE TPAMI, Nature Machine Intelligence, and Science Robotics has likewise been treated as eligible evidence. We document the invitation, the venue's selectivity, and the volume of review work because USCIS sometimes argues that peer review is expected of senior researchers and therefore not extraordinary. Whether a given pattern of review service is sufficient depends on the venue, the volume, and the broader record.
This is where AI-researcher RFEs most commonly land, and where careful petition design matters most. Citations alone are typically not sufficient. Stronger cases tend to show independent adoption of the work: third-party labs replicating or building on the contribution, open-source repositories integrating the model or method, industry products incorporating the research, or standards bodies referencing the work. We separate citations from the petitioner's collaborators and former co-authors from genuinely independent citations, because USCIS officers often do this analysis themselves. Quantifiable industry deployment — model download counts, product integration, platform incorporation — has been among the cleaner forms of this evidence in past cases. Whether any specific record clears the "major significance" threshold is ultimately a discretionary determination.
Top-tier AI conference papers have generally been treated as scholarly articles for EB-1A purposes, though this is not codified by regulation and individual officers may push back. Workshop papers can count but typically receive less weight unless the workshop itself is competitive and well-recognized. arXiv preprints alone have generally not been treated as scholarly articles in this regulatory sense — peer-reviewed publication or acceptance at a peer-reviewed venue is the more common threshold. Authorship order matters less than USCIS sometimes implies: in many AI subfields, last-author or senior-author position carries more signal than first-author for established researchers, and we explain that convention in the petition rather than letting an officer apply biology-publishing assumptions to a CS context.
This criterion rarely fits pure AI research. It may have application when applied AI work is shown at SIGGRAPH, CHI, or comparable venues with juried exhibition components, or when generative-art or interactive-AI work is exhibited at curated venues. We use this criterion only where there is a genuine fit and prefer comparable-evidence framing where the fit is strained.
Senior IC roles at industry research labs (for example, Google DeepMind, Meta FAIR, Microsoft Research, Anthropic, OpenAI, Apple ML Research, Amazon Science, NVIDIA Research), principal investigator status on grants, technical lead positions on production AI systems, lab head or research director roles, and named-chair faculty positions have supported this criterion in past cases. We document organizational distinguished-ness with recognition data — funding history, alumni outcomes, publications, hiring competitiveness — and explain the criticality of the specific role rather than relying on title alone. Whether an organization is sufficiently distinguished and a role sufficiently leading or critical is a fact-specific analysis.
Senior research scientist compensation at major AI labs often sits well above prevailing wage benchmarks, but the salary criterion is not a foregone conclusion even at high comp levels. We file this criterion with the petitioner's full compensation package — base, target bonus, vested and unvested equity — compared against OFLC, BLS, and industry-specific datasets such as Levels.fyi for the same role and seniority. Because equity is typical in this field, we also explain how restricted stock units convert to expected cash value, and we document the comparison against multiple benchmarks rather than relying on any single source. Whether a particular comparison persuades USCIS depends on the data sources and the framing.
Does not apply to AI research.
These objection patterns are decoded in the RFE Encyclopedia →
You'll hear back within 48 hours whenever a petition is being drafted or an RFE is on the clock. No ghosting.
Every petition is drafted from a fresh interview-extracted fact sheet. We don't recycle petitions or rec letters across unrelated clients.
We file on every criterion we can credibly defend. When a criterion is thin, we fold it into "Original Contributions of Major Significance" rather than stand it up as its own weak argument.
RFE response is a separate flat fee of $2,000 to $5,000, quoted before any work begins. Strategy consultations, whether-to-respond conversations, and post-denial planning are not billed hourly.
We use a structured SOAR (Situation, Obstacle, Action, Result) interview process to understand the client's actual work, including in technical and niche fields where the record doesn't speak for itself.
We draft reference letters from the interview and evidence review — included in the petition fee — then coordinate with recommenders for signature. We don't leave recommenders to produce their own letters.
RFEs aren't surprises. Every petition is drafted with our standing RFE response framework in mind so that if an RFE lands, we're executing a plan, not starting from scratch.
The initial call is a candid read on whether the case is defensible — not a pitch. If we think the profile doesn't support EB-1A right now, we'll tell you.
Tell us about your immigration needs and we'll be in touch to discuss how we can help.
Immigration counsel to Fortune 500 employers at a national firm · Adjudicated 12,000+ visas at the U.S. Consulate, Mexico · Working in U.S. immigration since 2008
Featured in Newsweek, Condé Nast Traveler, Daily Mail