No university ranking measures how good an artificial intelligence degree is to study. The AI-related rankings in wide use count different things: faculty papers at selective conferences, papers in top journals and conferences by subject, reputation surveys, and citation data. Two of them can place the same school far apart without either being wrong, so the useful question is what each list measures and whether that matches what an applicant needs.
Key takeaways
- CSRankings is metrics-only. It counts publications by faculty at selective conferences in areas such as AI, machine learning, vision, and language, using DBLP data (September 2026 on the site).
- ShanghaiRanking's Global Ranking of Academic Subjects lists Artificial Intelligence among 57 subjects and measures it mainly through papers at top conferences.
- Times Higher Education's overall 2026 method gives reputation surveys 33 percent of the total score. Its subject tables reuse the method with adjusted weights, and the subject page does not name an AI table.
- All of these rate research strength. None rates teaching quality, class size, tuition, or outcomes in a specific master's program.
- Changing the weights can reverse the order of two schools, as the illustrative table below shows.
Five lists, five different questions
Each organization picks inputs it can collect for many universities at once. That is why the lists favor research output, which is easy to count, over teaching, which is not. The table summarizes what the sources reviewed say each one measures. Weights are given only where the publisher prints them in readable form.
| Ranking | What it counts | AI-specific table? |
|---|---|---|
| CSRankings | Faculty publications at selective conferences, by area; no surveys and no citation counts | Yes, as AI, machine learning, vision, language, and web areas |
| ShanghaiRanking GRAS 2025 | Faculty awards and editors, top-venue papers, journal quality, citation impact, international collaboration | Yes, as one of 57 subjects |
| Times Higher Education 2026 | Teaching, research environment, research quality, international outlook, industry | Computer science table; no AI table named on its subject page |
| QS subject rankings | Academic and employer reputation surveys, citations per paper, h-index, international research network | Data Science and AI subject |
| U.S. News Best Global Universities | Bibliometric indicators built from Clarivate's Web of Science data | Artificial Intelligence subject |
CSRankings: a publication count with a short list of venues
The CSRankings site describes itself as a metrics-based ranking of computer science institutions built from faculty publications at selective conferences. The AI-related areas it lists are artificial intelligence (AAAI and IJCAI), computer vision (CVPR, ECCV, ICCV), machine learning (ICLR, ICML, NeurIPS, KDD), natural language processing (ACL, EMNLP, NAACL), and web and information retrieval (SIGIR, WWW). Its project documentation says it deliberately avoids surveys and also avoids citation counts, which it calls easy to manipulate, and that its faculty lists cover tenure-track faculty only. A school with a strong teaching-focused master's program but few listed faculty will therefore score low, and a school whose faculty publish at other venues will not be credited for them.
ShanghaiRanking: AI measured through conferences
The methodology page for the 2025 Global Ranking of Academic Subjects groups its indicators into five categories: world-class faculty (laureates, highly cited researchers, journal editors, leaders of academic organizations), world-class output (top-venue papers and awards), high-quality research (papers in top-quartile journals), research impact (category-normalized citations), and international collaboration. Each score is scaled against the top institution, and the square root of that percentage is multiplied by the indicator weight. For AI, the top-venue indicator uses conference papers, and the page states that 26 top conferences are used for computer science and engineering and for AI. A university must also pass a subject-specific publication threshold for 2020 to 2024. The weights table did not display as text on the page, so this article does not quote them.
Times Higher Education, QS, and U.S. News
For its 2026 world rankings, Times Higher Education weights teaching at 29.5 percent, research environment at 29 percent, research quality at 30 percent, international outlook at 7.5 percent, and industry at 4 percent. Inside those groups, teaching reputation carries 15 percent and research reputation 18 percent, so a third of the score comes from opinion surveys. The subject tables, including computer science, use the same method with weights adjusted for each field.
QS describes its subject rankings as using up to five indicators (academic reputation, employer reputation, citations per paper, h-index, and international research network), with a separate weighting for each subject and a minimum publication count. U.S. News, whose 2026-2027 subject rankings use Clarivate's Web of Science data, says it set indicator weights by its own judgment after consulting Clarivate, log-transforms skewed indicators, and scales scores so the top institution gets 100. Neither publisher's full weight table for its AI subject was readable in the sources opened, which is itself a fact worth knowing: an applicant often cannot reconstruct the order from the published inputs.
Same schools, different weights: an illustrative example
The numbers below are invented to show the mechanism and do not describe any real institution. School A has a stronger research-output score, and School B has a stronger reputation score.
| Output score | Reputation score | Total, output weighted 70% | Total, reputation weighted 70% | |
|---|---|---|---|---|
| School A | 90 | 60 | 81.0 | 69.0 |
| School B | 70 | 85 | 74.5 | 80.5 |
School A leads under the first weighting and School B under the second. Real lists have more indicators, but the logic is the same: the order depends on a choice that readers rarely see.
Checking a program directly
- Read the program's own page for required courses, credit hours, length, and total cost, and note the academic year of the cost figure.
- Look up named faculty in the research areas of interest, and check where they publish.
- Confirm the school's recognition and any program-level accreditation with the official bodies that issue them, not with a ranking page.
- Ask the admissions office for graduate outcomes and for how many students receive funding.
- Check the edition and date of any ranking cited, since weights and thresholds change between editions.
FAQ
Is there an official ranking of AI universities in Canada or the United States?
No government agency in either country publishes one. The lists discussed here are produced by private organizations, a university-based project, or a media company.
Why does a school rank high on one AI list and low on another?
The lists count different things. A ranking built on conference papers rewards research output, while one with large survey components rewards reputation, and a school can be stronger on one than the other.
Does a high research ranking mean a better master's program?
Not necessarily. Research rankings measure faculty output, not teaching, course design, or graduate outcomes in a given program.
Which ranking should an applicant trust?
None should be used alone. Each is useful for the question it answers, and the program's own pages and official data answer the questions rankings do not.
Related reading
- AI programs in the USA and how to choose
- Ranking claims in a national AI program
- Why version and date matter in comparisons
- Measuring a result against a benchmark
More in AI Basics.
Last verified: October 2026
Sources
- CSRankings: Computer Science Rankings, Emery Berger and contributors, data dated September 2026
- CSrankings project documentation (README), GitHub, 2026
- Global Ranking of Academic Subjects 2025 Methodology, ShanghaiRanking Consultancy, 2025
- World University Rankings 2026 methodology, Times Higher Education, 2025
- World University Rankings by Subject 2026, Times Higher Education, 2026
- QS World University Rankings by Subject methodology, QS Quacquarelli Symonds, 2026
- How U.S. News Calculated the 2026-2027 Best Global Universities Subject Rankings, U.S. News and World Report, 2026