EB-2 NIW for AI Engineers and Data Scientists: Evidence Guide

AI is a national priority, but that alone will not win an NIW. See how machine learning engineers and data scientists prove national importance in 2026.

IMMIGRATION INSIGHTS

10/22/20265 min read

Artificial intelligence sits near the top of US policy priorities. It appears on the federal list of critical and emerging technologies, the White House has published an AI Action Plan, and USCIS guidance names AI as an area where national importance can be shown. So why are so many NIW petitions from machine learning engineers and data scientists denied?

The answer is that USCIS evaluates your specific proposed endeavor, not the importance of AI in general. This guide explains how USCIS approaches AI and data science petitions, the evidence that works for industry engineers and researchers, and the mistakes that lead to denials.

Key takeaways

USCIS guidance treats an advanced STEM degree, especially a PhD, combined with work in a critical and emerging technology such as AI, as a strong positive factor. A degree alone is not enough.
National importance comes from your endeavor's impact, not from articles about the importance of AI.
Industry engineers without many publications can rely on patents, open-source adoption, deployed systems and technology licensed or adopted beyond their employer.
The AAO has denied AI petitions where benefits were limited to one employer or where citations and letters lacked specifics.
Letters from interested US government agencies carry particular weight under USCIS guidance.

How USCIS views AI and data science petitions

The USCIS Policy Manual includes guidance specifically for petitioners in science, technology, engineering and math. It recognizes that an advanced STEM degree, particularly a PhD, can be a strong positive factor when combined with other evidence, especially for work tied to critical and emerging technologies. It cites "certain critical areas of artificial intelligence" as an example and refers to the federal Critical and Emerging Technologies list.

In August 2026, the White House released an updated version of that list as part of its national security science and technology strategy. AI remains on it, with emphasis on areas such as foundation models, interpretability, adversarial robustness and AI security, and privacy-preserving machine learning. If your work falls in one of these areas, cite the current list and explain precisely how your work fits.

USCIS also highlights positive indicators for technology work, such as widespread interest in adopting or licensing your technology and evidence that it has influenced similar technology at other companies.

Why generic AI arguments fail

In a March 2026 non-precedent decision, the AAO dismissed the appeal of a machine learning engineer working on edge AI. The petitioner submitted general articles on edge AI, a consulting firm report and a White House fact sheet. The AAO found these did not show the national importance of the petitioner's specific endeavor, emphasized that benefits to a specific employer alone are not enough, and noted that STEM considerations do not replace the national importance requirement.

USCIS guidance offers a related example of work that falls short: a software engineer adapting an employer's code for its clients. The work is skilled, but its benefits stay within one company and its customers.

The lesson is consistent with Matter of Dhanasar: define a proposed endeavor with broader implications, then prove it with evidence about your work.

Defining a strong AI proposed endeavor

Weak: "I will advance artificial intelligence in the United States."

Stronger: "I will develop and deploy privacy-preserving machine learning methods for medical imaging that allow hospitals to train diagnostic models without sharing patient data, building on methods I developed that are now used by three hospital research networks."

A strong endeavor names the problem, the method, who benefits beyond your employer, and the evidence that you have already made progress. Connect it, where accurate, to documented national priorities such as cybersecurity, healthcare, critical infrastructure, energy or scientific discovery.

Evidence for AI engineers and data scientists

If you are a researcher

Publications in respected AI venues, with citations by independent researchers. Separate self-citations and compare your numbers with peers.
Examples of other researchers building on or benchmarking against your work.
Datasets, models or code released with your papers and adopted by others.
Peer review for conferences and journals, invited talks and competitive grants.

If you work in industry

Many strong AI petitioners have few publications. They rely instead on evidence of real-world adoption:

Patents, especially those licensed, cited by other patents or used in products.
Open-source projects with documented adoption: downloads, stars, forks, dependent projects and contributors from other organizations.
Deployed systems with measurable reach: users served, transactions processed, error rates reduced, or costs saved for customers across an industry.
Technical leadership, such as designing a system that other teams or companies adopted.
Industry recognition, such as invited conference talks, technical awards and trade press coverage.

Document your role in each achievement with employer letters that describe your specific contributions, not only the team's.

Support from independent experts and agencies

Recommendation letters should come from independent experts who can explain specifically what you built and how it influenced the field. USCIS guidance gives particular weight to letters from interested US government agencies, so if your work supports a federal program or agency mission, a letter from that agency can be valuable.

Common mistakes in AI NIW petitions

Relying on reports about the AI market instead of evidence about your work.
Describing an endeavor that is really your current job description.
Presenting citation counts without context, or counting citations that arrived after filing. Eligibility must exist when you file.
Letters with general praise and no specific examples.
No employer documentation of your individual role in team projects.

In an April 2025 decision, the AAO found a computer vision engineer met prong 1 but failed prong 2, pointing to modest citation counts without self-citations separated, letters without specific examples and a lack of employer documentation of her role. Since August 5, 2026, USCIS may also deny a petition lacking required initial evidence without first issuing a request for evidence. See our NIW approval rate analysis.

NIW or EB-1A for AI professionals?

If you have major recognition, such as highly cited papers, significant awards, judging experience, media coverage and a leading role at a distinguished organization, EB-1A may fit. Our guide to EB-1A for software engineers and our post on original contributions of major significance explain that standard. Many AI professionals file NIW first, or file both.

With the H-1B now using a wage-weighted lottery and subject to a contested $100,000 payment for some petitions, a self-petition can also reduce dependence on a sponsor. See H-1B $100,000 Fee: Are NIW and EB-1A Better Alternatives?

HoatPen writes NIW and EB-1A petitions for AI, machine learning and data science professionals, translating technical work into a clear national importance argument. Book a 1-hour profile consultation to review your record. The $100 fee is credited toward your writing package if you move forward.

Frequently asked questions

Can a machine learning engineer get an NIW?

Yes, if the engineer qualifies for EB-2 and shows a specific proposed endeavor with national importance, plus evidence of being well positioned to advance it. Work limited to one employer's products usually needs a broader framing and evidence of impact beyond the company.

Do I need publications for an NIW in AI?

No. Publications help, but industry engineers can rely on patents, open-source adoption, deployed systems, licensing and independent expert letters to show their impact.

Is AI on the critical and emerging technologies list?

Yes. AI appears on the federal Critical and Emerging Technologies list, which was updated in August 2026. USCIS guidance cites critical areas of AI as an example of work that can support national importance.

Is data science an NIW-eligible field?

Data science can support an NIW when the proposed endeavor applies it to a problem of national importance, such as public health, cybersecurity, energy or financial system stability, and when the evidence shows impact beyond one employer.

Can I use an NIW instead of an H-1B?

An NIW is a green card petition you file yourself, so it does not depend on an employer or the H-1B lottery. It does not provide work authorization while the I-140 is pending, so many people keep a temporary status in the meantime.

HoatPen is a technical writing and consulting service, not a law firm. This article is general information and is not legal advice. Confirm your eligibility and the current rules with an immigration attorney before filing.

Navigation

admin@hoatpen.com

+1 (640) 344-3582

Disclaimer

HoatPen ~ part of HoatGroup LLC, is a technical writing and consulting service, not a law firm. We do not provide legal advice or representation in any jurisdiction. Our services focus on helping clients prepare high-quality written materials, including petitions, recommendation letters, and supporting documents based on publicly available immigration guidelines and professional writing standards.