If you do AI research and someone just floated the EB-1A green card for AI researchers as an option, the first thing to get straight is what it actually is. The EB-1A is not a work visa. It is a category of employment based permanent residence, the "extraordinary ability" green card, and it sits in a different part of immigration law than the temporary visas most engineers hear about first. It shares some logic with the O-1, but the two solve different problems. Nothing here is legal advice, just a map of the category.
EB-1A vs the O-1: permanent status versus a renewable visa
The O-1 is a nonimmigrant visa. It lets you work in the US for a defined period, usually up to three years at a time, and it eventually needs renewal or conversion into something else. The O-1 visa page from USCIS lays out that structure. The EB-1A sits in a different category: it is one of the first preference employment based green card routes, meaning approval leads to permanent residence rather than a renewable stay.
Plenty of researchers file both, in sequence or close together, since the O-1 buys time while the slower green card case works its way through. Related reading: the O-1 visa for AI researchers and engineers covers that side in more detail.
You can self-petition, no employer required
Most employment green cards need a US employer to sponsor you and, for many categories, labor certification (PERM), the process of showing no qualified US worker was available for the role. The EB-1A skips that. Under the EB-1 first preference category described by USCIS, a person with extraordinary ability can file the petition on their own: no job offer, no employer letter, no PERM. That is why the category draws attention from independent researchers, people between jobs, and founders without a conventional sponsor. Self-petitioning does not lower the bar; it just means the evidence has to carry the whole case, with no employer vouching alongside you.
The evidence criteria, in plain language
USCIS evaluates EB-1A cases against a defined list of criteria described in the USCIS Policy Manual chapter on extraordinary ability. The regulation lists ten categories, though two rarely apply to AI research (major awards and commercial success in the performing arts). The commonly cited approach is to meet at least three of the remaining criteria, or show a single major, internationally recognized one time award that stands in for the whole list on its own. Read these as prompts for self reflection, not a scoreboard.
- Awards for excellence in the field, recognized nationally or internationally, such as a best paper award at a major conference.
- Membership in associations that require outstanding achievement to join.
- Published material about you, press or trade coverage discussing your work specifically, not a passing credit.
- Judging the work of others, reviewing papers, serving on a program committee, judging a competition.
- Original contributions of major significance, work shown to have changed practice in the field, usually evidenced through citations or adoption.
- Authorship of scholarly articles in professional journals or other major media.
- A leading or critical role at an organization with a distinguished reputation.
- A high salary relative to others in the field, shown with comparative data.
Meeting three of these is a threshold question, not the end of the review. USCIS then weighs the totality of the evidence, so three thinly supported categories do not fare the same as two or three backed by strong documentation.
A self-assessment for AI researchers with papers and citations
Most people out of a serious PhD program, a known lab, or a senior applied research role already have a few of these categories half built without realizing it. A handful of well cited papers cover original contributions and scholarly authorship at once. Reviewing for a workshop covers judging. A senior title at a recognizable lab can support a critical role claim, and public compensation data can support the salary criterion. Engineers without publications can often build a case around a critical role, salary evidence, and one strong original contribution, such as an open source project with documented adoption.
Do the inventory honestly before paying anyone. List your papers and citation counts, awards or fellowships, reviewing or judging work, your pay against public benchmarks, and anything you built that people actually use. Write one sentence under each item on why it matters, not merely that it happened. That tells you more than a generic checklist ever will.
A rough gut check: if you had to convince three people who work in your subfield, separately, that you belong among the strong researchers in it, and you can point to something specific for each of them, a paper they would recognize, a talk, a system they have used, you likely have the start of a case worth taking to an attorney.
Common misconceptions worth clearing up
The biggest one is that EB-1A is reserved for people with a Nobel Prize or a household name. Plenty of working researchers qualify who would never call themselves famous outside their own subfield. The second is assuming a PhD alone is enough; it helps because it tends to come with publications, but the degree itself is not one of the criteria. The third is treating a denial as final. Evidence gaps get identified, and petitioners often refile with a stronger record. Outcomes vary with how a case is documented and who adjudicates it, so nothing here is a guarantee.
The honest takeaway
The EB-1A is a real, workable path to a US green card for AI researchers whose profile includes publications, citations, recognition, or a clearly documented critical role, and self-petitioning removes the employer dependency that trips up other categories. It is evidence based work, not a formality, and it is not guaranteed no matter how strong the resume looks on paper. Do the self-assessment first, be honest about the gaps, and talk to an immigration attorney who files these cases regularly before spending money on the process. This article is informational only, not legal or immigration advice; requirements change, so verify current rules with USCIS's EB-1 page or a licensed immigration attorney before making any decisions.
If the green card question is one piece of a bigger decision about where to build a research career, look at the whole board rather than one category alone. The AI Relocation Guide lays out visa and permanent residence routes, after tax pay, and cost of living across 21 countries, so you can compare all 21 countries before betting a career on a single immigration category.
