Separate the term “AI” from the careers it encompasses
Job postings use vague terminology. Under the “AI” label, there are at least four families of roles, with job titles that vary from one company to another.
- Research: designing new machine learning methods, most often with a doctorate.
- Machine learning engineering: training, evaluating and deploying models in products.
- Data engineering: building pipelines that collect, clean and store data.
- Data analysis and data science: interpreting data to support business decisions.
Other roles surround these: product management, security, compliance and technical writing. Before choosing a study pathway, it is useful to read several actual job descriptions and identify the skills required, rather than relying on the job title alone.
Assess demand with caution
The World Economic Forum’s (WEF) Future of Jobs 2025 report surveys more than 1,000 employers. According to this edition, 86% of respondents expect AI and information processing to transform their business by 2030. AI and machine learning specialists are among the three fastest-growing roles in percentage terms, alongside big data specialists and financial technology engineers.
Among the fastest-growing skills, the report ranks AI and big data first, followed by networks and cybersecurity, then technological literacy. According to the report’s analyses, AI and information processing could create around 11 million jobs and displace around 9 million by 2030.
These figures describe employers’ expectations, not job openings. Above all, they point to a shift: some tasks are disappearing, while others are emerging.
Map out study pathways
There is no single educational route into these careers. Depending on the role sought, common programs include computer science, mathematics and statistics, software or electrical engineering and, more recently, programs dedicated to data science or AI.
Research roles generally require a doctorate. Engineering roles are often accessible with a strong bachelor’s degree in computer science, supplemented by a master’s degree or projects. Analysis roles accept more varied backgrounds, including economics, social sciences or management, combined with a foundation in programming.
Almost all programs require mathematics prerequisites: linear algebra, probability, statistics and calculus. Before choosing a country, it is useful to compare these prerequisites, required courses and the scope for practical projects. Degree titles sometimes conceal very different content.
Read the QS 2026 ranking in data science and AI
The QS subject rankings (QS World University Rankings by Subject), published on March 25, 2026, include a “Data Science and Artificial Intelligence” table. This table is new: it first appeared in 2023 with 20 universities and has grown every year since. The 2026 edition ranks 201 universities, compared with 104 in 2025. QS identifies it as the subject with the most new entries.
This growth explains shifts in rank that do not always reflect a change in quality. When more than a hundred institutions enter the table at once, those already listed can fall without having changed.
| QS 2026 rank | University | Country or territory |
|---|---|---|
| 1 | Massachusetts Institute of Technology | United States |
| 2 | Stanford University | United States |
| 3 | National University of Singapore | Singapore |
| 4 | Nanyang Technological University | Singapore |
| 5 | Carnegie Mellon University | United States |
| 8 | Harvard University | United States |
| 9 | University of Cambridge | United Kingdom |
| 10 | Tsinghua University | Mainland China |
The two institutions tied for 6th place are not reproduced here. QS notes that the National University of Singapore is the highest-ranked institution outside the United States, and that Tsinghua appears in the table for the first time.
Place the table within the overall QS rankings
The AI table is just one of the 55 subject rankings in the 2026 edition. According to Istanbul Technical University, citing QS figures, 6,273 universities were analysed and 1,908 were ranked across all subjects. The number of institutions varies considerably from one table to another, from a few dozen to several hundred.
QS applies thresholds: an institution that does not meet the minimum level of research output or reputation in a subject is not included. Not appearing in a table therefore does not mean that an institution does not offer the program, only that it has not been ranked.
This caution also applies to the neighbouring subject, computer science and information systems, which QS describes as one of the most extensively ranked in the world. For the same institution, rankings in computer science and data science can differ significantly because the groups of institutions being compared are not the same.
Check the criteria behind the rank
In its subject rankings, QS combines academic reputation, employer reputation, citation impact and the h-index. These indicators primarily measure research and reputation. They say little about the quality of first-year teaching, class sizes or access to internships.
For a bachelor’s degree, more concrete information often carries greater weight: course content, internship opportunities, laboratories accessible to students and placement rates published by the institution. For a doctorate, the profile of the research supervisor and laboratory matters more than the university’s rank.
A new and rapidly growing ranking is best cross-checked against other sources: program websites, researchers’ publications, alumni accounts and lists of industry partners.
Compare university degrees and short training programs
There are many AI certificates and intensive training programs, and their promoters tout their results. Among the sources consulted for this article, no published, comparable data on the employment outcomes of these short training programs was found. It is better to stick to what can be verified.
The World Economic Forum ranks AI and big data skills as the fastest-growing skills: demand is for skills, not for a particular type of degree. However, research roles remain linked to graduate degrees, and some regulated professions require a recognized degree.
To compare a short training course with a university program, the criteria published by the institution are the most useful: detailed course content, duration, the inclusion of assessed projects, recognition of credits by a university and a list of employer partners. Theoretical foundations in mathematics remain useful as tools change.
What about Canada?
Canada launched its Pan-Canadian Artificial Intelligence Strategy in 2017, which the Canadian Institute for Advanced Research (CIFAR) describes as the world’s first national AI strategy. It is built around three national institutes: Mila in Montréal, the Vector Institute in Toronto and Amii in Edmonton. These institutes are linked to nearby universities and host researchers and graduate students.
No specific Canadian rank is cited here for the QS 2026 subject table: the detailed ranking could not be consulted in full at the time of writing and can be checked on the QS website.
To explore study pathways, StudyCanada offers the University of Toronto profile, the Université de Montréal profile, the B.Sc. in computer science page at McGill University and the computer science program page at the University of Waterloo. Each profile specifies mathematics prerequisites and requirements for international students.