Blog post
AI and inequality in higher education
AI adoption in higher education is often described as a question of innovation or modernization. But it is equally a question of equity: when better-resourced institutions are the first to build AI programs, create governance structures, and attract partnerships, they gain further advantages in visibility, funding, staff recruitment, research performance, and student opportunity. Regional and less-resourced universities, by contrast, may lack the financial, technical, and organizational means to develop equivalent responses.
This creates a risk of system stratification between flagship and regional universities, research-intensive and teaching-focused institutions, well-funded and less-resourced institutions, and institutions close to policy networks and those at the margins, as described in the three country cases described below.
Canada: decentralized governance, concentrated leadership
Canada’s AI governance environment is decentralized and multi-layered. Universities operate within a policy landscape shaped by federal and provincial actors, public guidance, sectoral expectations, and research funder requirements. Within higher education, the most coordinated AI-related activity appears among leading research-intensive universities. Shared frameworks and collaborative discussions are being driven primarily by major institutional networks, and top universities are more likely to have established formal governance arrangements such as central task forces, advisory bodies, and university-wide guidance, compared to lower-tier institutions that have fewer resources to respond.
China: common expectations, different implementation depth
China offers a more centralized and compliance-oriented approach. Universities are increasingly expected to regulate AI use across teaching, assignments, examinations, thesis writing, research, and data governance. Institutional policies emphasize disclosure, academic integrity, data security, and ethics review. However, common rules do not guarantee equal implementation. Leading universities appear to be the first to translate national expectations into detailed procedures, templates, review arrangements, and enforcement systems. Institutions with weaker administrative or technical capacity may comply formally but lack the same depth of institutional readiness.
Kazakhstan: national ambition, visible system gaps
Kazakhstan shows most clearly how AI adoption can reinforce institutional inequality. National policy has moved quickly, with ambitious initiatives aimed at building AI skills, innovation capacity, and a broader ecosystem for digital development. Yet universities differ sharply in AI programs, research capacity, and policy development. A small number of leading institutions have stronger academic offerings, dedicated research centers, and clearer guidance for AI use. Many regional universities remain at a much earlier stage, with limited infrastructure, fewer specialized programs, and weaker governance arrangements, demonstrating that strong national ambition does not automatically produce balanced system-wide implementation.
Main policy insights
The evidence from all three countries points to a common conclusion: AI adoption tends to advance fastest where institutional capacity is already strongest. For educational planners, this raises an important concern: if AI policy focuses only on innovation and uptake, it may unintentionally deepen inequalities across the higher education sector. Equity must therefore become a central consideration in AI governance. As a result, five priorities stand out for ministries, planners, and higher education leaders:
- Treat AI adoption as a system equity issue, not only a technology issue;
- Combine national direction with institutional capacity-building;
- Avoid relying exclusively on flagship universities as the drivers of reform;
- Invest in shared tools and support mechanisms that benefit the whole system;
- Measure success not only by innovation outputs, but by inclusive uptake across institutions.
More specific policy options are described below:
-
Establish minimum system-wide standards: Governments and higher education authorities should define baseline expectations for AI use in teaching, research, administration, ethics, and data protection, helping to ensure that responsible practice does not depend entirely on the internal resources of individual institutions.
-
Build shared support structures: National or sector-wide support mechanisms can help reduce inequality. These may include model policies, guidance materials, staff training, shared ethics review tools, and common templates that institutions can adapt locally.
-
Target support to less-resourced institutions: Regional and less-resourced universities need dedicated funding, technical assistance, and professional development if they are to participate meaningfully in AI adoption.
-
Promote collaboration across the whole system: Governments should encourage peer-learning networks, inter-institutional partnerships, and resource-sharing arrangements that include the full higher education sector, not only elite universities.
-
Monitor distributional effects: AI policy should be evaluated not only by levels of uptake, but also by equity. Policymakers should track which institutions have governance frameworks, trained staff, research capacity, and infrastructure, and which do not.
Acknowledgement
This project was supported by the Swiss foundation Movetia. The sponsor influenced neither the research design nor the interpretation of the results.
Reference
Denisova-Schmidt, E., Kuzhabekova, A., & Xu, Z. (2026). Campus AI governance: What Canada, China, Kazakhstan reveal. University World News, 15 April, 2026. Campus AI governance: What Canada, China, Kazakhstan reveal.
