EB-1A AI Expert Opinion Letter
Independent expert opinion letters for EB-1A Extraordinary Ability petitions in artificial intelligence, machine learning, data science, and enterprise AI. I have prepared 20+ expert letters for AI professionals, and I only write one when the evidence supports it.
What an EB-1A expert opinion letter is
An EB-1A expert opinion letter is an independent written evaluation by a recognized specialist in the petitioner's field, submitted as supporting evidence in an Extraordinary Ability petition. It explains what the petitioner built, why that work was original, who adopted it, and how it compares to the state of the art, in language an adjudicator who is not an AI practitioner can evaluate.
It is not a recommendation letter and it is not legal argument. It is a technical assessment of the record by someone with the standing to make it.
The ten EB-1A criteria and where an expert letter helps
A petitioner must show a one-time achievement of major international recognition, or satisfy at least three of the ten criteria at 8 CFR 204.5(h)(3). The table below maps each criterion to the evidence AI petitioners typically hold, and flags where an independent expert opinion carries the most weight.
| Criterion | Typical evidence in AI cases | Expert letter |
|---|---|---|
| Nationally or internationally recognized awards for excellence | Best-paper awards at recognized ML venues, industry AI awards, internal awards only where selection is competitive and externally recognized. | Supporting |
| Membership in associations requiring outstanding achievement | IEEE senior membership, ACM distinguished status, invitation-only technical bodies with judged admission. | Supporting |
| Published material about you in professional or major trade media | Trade press covering your system or product, technical interviews, analyst reports that name you. | Supporting |
| Judging the work of others | Peer review for ML conferences and journals, hackathon and grant judging, standards committee review, technical due diligence panels. | Supporting |
| Original contributions of major significance | Novel architectures or methods, systems adopted beyond your employer, open-source projects with independent adoption, patents that others build on. | Core |
| Authorship of scholarly articles | Peer-reviewed papers, cited preprints, substantive technical publications in recognized venues. | Supporting |
| Display of work at artistic exhibitions | Rarely applicable to AI petitioners. | Not typical |
| Leading or critical role for distinguished organizations | Ownership of an AI system central to the organization's business, founding or leading an AI function, named technical authority on a flagship platform. | Core |
| High salary or significantly high remuneration | Compensation benchmarked against role and geography for AI and ML positions. | Supporting |
| Commercial success in the performing arts | Not applicable to AI petitioners. | Not typical |
Meeting three criteria does not end the analysis. Following Kazarian v. USCIS, adjudication is a two-step process: first counting which criteria are satisfied, then a final merits determination on whether the record as a whole shows sustained acclaim and places the petitioner among the small percentage at the very top of the field. An expert opinion letter is often most valuable at that second step, where the question is qualitative rather than a checklist.
What I address in an EB-1A letter
- The petitioner's original contributions to AI or machine learning, described precisely enough that a non-specialist can see what was new
- Independent adoption: who outside the petitioner's own employer uses the work, and what changed because of it
- How the contribution compares to prior and contemporaneous work in the same subfield
- The petitioner's role and level of responsibility relative to the organization's technical output
- Practical and commercial impact, quantified where the record supports it
- Recognition from parties with no stake in the petition
- My own basis for making the assessment, including why I am qualified to evaluate this specific area of AI
Where AI EB-1A records are usually weakest
Patterns I see repeatedly when reviewing AI petitions. None are fatal, but each is easier to address before filing than after a Request for Evidence.
Impact described only inside the employer
"Deployed to production and used by every team at the company" establishes seniority, not field-wide significance. The criterion asks about contributions to the field. Evidence of adoption, citation, or replication by parties with no commercial relationship to the petitioner is what moves the analysis.
Confidential work with no public trace
A great deal of the strongest applied AI work sits behind NDAs. This is a genuine problem, not a disqualifying one. Independent expert evaluation exists partly to give confidential work an evidentiary footing, but it needs something to stand on: internal metrics, architecture documents, patents, or named attestations from people who can speak to it.
Citation counts treated as the only measure
Applied AI practitioners frequently have modest citation records and enormous production impact. A record built solely around a citation table often understates an industry petitioner. The reverse is also true: a strong citation record with no explanation of what the work enabled leaves the adjudicator to guess.
Generic contribution language
"Pioneered cutting-edge deep learning solutions" is unfalsifiable, and adjudicators discount it. Specific claims, tied to artefacts, are what an expert can actually corroborate.
What an expert letter cannot do
It cannot substitute for evidence that does not exist, argue the legal standard, decide eligibility, or guarantee an outcome. It cannot make a record extraordinary if it is not. I am not an attorney and I do not provide legal advice; the letter is one input to a petition that your immigration counsel builds and argues.
Process and turnaround
- Initial review - you or your attorney send the CV, petition context, and the evidence you plan to rely on
- Evidence assessment - I read the record and tell you candidly whether it supports a strong expert opinion
- Drafting - if it does, I prepare the letter, typically in 5–10 business days, measured from receipt of complete materials, not from first contact
- Attorney coordination - I work with counsel so the letter addresses the right technical questions without straying into legal argument
- Delivery - the signed letter goes to your attorney for inclusion in the petition
Expedited review in 2–3 business days is available where a filing or response deadline requires it. Full detail on the process and turnaround page.
Frequently asked questions
▸ Can an AI expert write a letter for an EB-1A petition?
Yes. An independent expert in artificial intelligence can provide an opinion letter evaluating a petitioner's original contributions, the significance of those contributions to the field, and their standing relative to peers. Expert opinion letters are routinely submitted as supporting evidence in EB-1A petitions, most often to substantiate the original contributions of major significance criterion and the leading or critical role criterion, and to support the final merits determination.
▸ How many EB-1A criteria do you need to meet?
USCIS requires evidence of a one-time achievement of major international recognition, or evidence satisfying at least three of the ten criteria at 8 CFR 204.5(h)(3). Meeting three criteria is not the end of the analysis. Under the two-step framework from Kazarian v. USCIS, the officer then makes a final merits determination on whether the evidence as a whole demonstrates sustained national or international acclaim and that the petitioner is among the small percentage at the very top of the field.
▸ Which EB-1A criteria does an expert opinion letter actually help with?
An expert letter is most useful for original contributions of major significance and for a leading or critical role, because both require an informed judgment about the field rather than a document that speaks for itself. An expert can explain why a particular architecture, system, or result mattered, who adopted it, and how it compares to prior work. Criteria such as awards, memberships, salary, and authorship are largely evidenced by documents directly, though an expert letter can still supply the context that makes their significance legible to a non-specialist reader.
▸ What qualifies as extraordinary ability in AI?
There is no AI-specific standard. USCIS applies the same regulatory criteria used for every field, so extraordinary ability in AI is demonstrated the same way as in any science: through original contributions that others build on, recognition from independent parties, critical roles at organizations with a distinguished reputation, and evidence of sustained acclaim. What varies is the evidence. In AI, impact often shows up as production systems, model adoption, open-source usage, and internal technical influence rather than only as citation counts.
▸ Does an expert opinion letter guarantee an EB-1A approval?
No. No letter, from any expert, guarantees any immigration outcome. An expert opinion letter is one piece of evidence within a petition assembled and argued by a licensed immigration attorney. Its value is that it lets an independent specialist explain technical significance that documents alone do not convey. Any provider promising an approval is describing something they cannot deliver.
▸ How is an expert letter different from a recommendation letter?
A recommendation letter is usually written by a colleague, manager, or collaborator who knows the petitioner personally and speaks to their character and work. An expert opinion letter is written by an independent specialist who has no prior relationship with the petitioner and evaluates the work against the state of the field. Adjudicators tend to weigh independent evaluations differently from letters by people with a personal or professional connection to the petitioner.
▸ What if the evidence does not support a strong opinion?
I will say so, before any letter is drafted. I charge for review time and preparation, and payment is never contingent on the content of the opinion or the outcome of the petition. If the record does not support a favorable expert opinion, telling you early is more useful than producing a letter that an adjudicator will discount.
Related
Request an EB-1A expert opinion letter
If you or your client is preparing an EB-1A petition in artificial intelligence, send me the record and I will tell you whether I can help.