Universities Are the Bottleneck

Ottawa’s AI for All strategy aims to accelerate AI adoption through infrastructure, tools, training, and talent development. Its ambitions are necessary, but universities will be central to whether they succeed.

Universities educate future workers, train researchers, and shape how new technologies are understood and used. Yet within academic culture, AI is still too often treated as a threat to manage rather than a capability to build.

If Canada wants AI for all, universities cannot remain a bottleneck.

Higher education institutions such as Western University will play a central role in whether Canada’s AI ambitions reach students, researchers, and the wider workforce. Photo by Josh Grignon.

POLICING AI IN THE CLASSROOM

Ottawa’s vision for empowering Canadians through AI extends beyond providing access to AI tools. Pillar 2 aims to build AI capability through literacy, practical skills, instructor training, and meaningful participation. Preparing the next generation is at the centre of that vision. 

Canadian universities educate nearly one-third of the population, preparing students to confront future problems that will increasingly require AI-enabled skills and solutions. Whether universities can fulfill that role depends on their willingness to adopt pedagogies that integrate AI. 

Some academics argue that AI has poisoned the learning environment and poses an existential threat to academia. They propose banning AI tools, prohibiting laptops and smartphones in class, blocking AI services on university networks, and returning students to handwritten examinations. 

Others go further, calling for AI to be abolished across entire disciplines and shunned in the classroom. This imagines the university as a space insulated from the technologies reshaping work and society. That is difficult to reconcile with preparing graduates for an AI-enabled future.

This defensive posture reflects a wider anxiety in higher education. Rather than teaching students to use AI responsibly, it seeks to remove the technology from the learning environment altogether. It has turned AI into something to police rather than something to teach. 

Pillar 2 sets out a necessary vision for AI literacy in Canada. But success will require more than training instructors and expanding access to AI tools. It will require universities to embrace AI as part of teaching and learning rather than treating it primarily as a problem to manage. 

Policy Action: Tie Public Funding to AI Integration 

Universities depend on public funding. While student enrolment accounts for roughly 30 percent of university revenue, federal and provincial support remains the largest single source at roughly 42 percent. That gives governments a legitimate role in shaping whether universities prepare students for an AI-enabled future. 

Governments should tie a portion of university funding to clear AI integration milestones through the performance agreements they negotiate with universities, such as Ontario’s Strategic Mandate Agreements. Universities should be required to show how AI literacy, responsible use, and practical skills are being built into curricula and degree requirements.

This cannot be limited to STEM. Many graduates build careers outside their field of study, where AI skills may be central to the work they do.

Integrating AI into university education will require clear policies, transparent expectations, and instructors willing to rethink how AI fits into teaching, assessment, and evaluation. Photo by Josh Grignon.

RESISTING AI IN RESEARCH 

Where Pillar 2 focuses on preparing the next generation to use AI, Pillar 4 turns to the researchers already shaping Canada’s knowledge economy. It aims to give them access to secure, sovereign AI tools and develop the talent needed to use them. But access means little when open opposition exists within academic culture.

The arts, humanities, and social sciences are no less important to Canada’s future than engineering, medicine, and technology. AI will reshape scholarship across every discipline. The challenge is that acceptance of AI is far less certain outside STEM.

Social scientists argue that AI distorts the conditions needed for meaningful research. They frame its use as incompatible with human inquiry and a threat to the integrity of qualitative scholarship. Refusing AI has become a statement about what it means to be a responsible researcher. 

These are not isolated cases. Over 400 researchers have signed a statement arguing that AI is inappropriate for all phases of reflexive inquiry. They reject AI on ethical grounds, arguing that its development is rooted in exploitative and colonial practices.

Pillar 4 sets out a necessary vision for AI-enabled research in Canada. But success will require more than sovereign infrastructure, talent development, and access to secure tools. It will require a research culture that sees AI as a legitimate tool for discovery rather than a threat to established ways of working.

Policy Action: Create Dedicated AI Research Funding 

AI will reshape scholarship across every discipline, but resistance is concentrated within the arts, humanities, and social sciences. If Canada hopes to build world-leading AI research capacity, it cannot afford for an entire segment of the research community to remain disconnected from these methods.

Governments should work with funding agencies like SSHRC to create dedicated streams for AI-enabled arts, humanities, and social science research. Existing competitions should remain intact, but new funding should reward projects that integrate AI methods and train graduate students in responsible use.

Funded projects should be required to document how AI was used, where its limits were encountered, and how researchers remained critically engaged with its outputs.

MAKING ROOM FOR CHANGE

Ottawa is investing in the infrastructure, talent, and tools needed for Canada to lead in AI. The next challenge is cultural. Universities must prepare students for an AI-enabled future and create research environments where new methods can be explored rather than dismissed.

The recommendations outlined here are intended to encourage that shift. For the strategy to succeed, universities must be willing to change with the technology.

AUTHORS’ NOTE

As social scientists who completed, or are completing, PhDs during the rapid emergence of AI, we approached these tools with curiosity. We tested where they helped, where they failed, and how they could be used responsibly, knowing that even exploring them could carry reputational risks.

Throughout our work, we encountered researchers arguing that AI has no place in qualitative inquiry, effectively calling into question the value of our own methods and findings. When one article made that case in absolute terms, we proposed a rebuttal grounded in our experience. It was rejected on the grounds that the debate had reached an acceptable conclusion, one that treated AI as an illegitimate research tool.

Such certainty is easier to maintain from the security of an established academic career. For students and early-career researchers, however, discouraging AI literacy may do more harm than good, regardless of discipline. As more graduates build careers beyond academia, AI skills are increasingly becoming part of the modern workplace. We have seen this expectation firsthand in our own job searches.

Integrating AI into university education will require clear policies, transparent expectations, and instructors willing to rethink how AI fits into teaching, assessment, and evaluation. Photo by Josh Grignon.

RESISTING AI IN RESEARCH 

Where Pillar 2 focuses on preparing the next generation to use AI, Pillar 4 turns to the researchers already shaping Canada’s knowledge economy. It aims to give them access to secure, sovereign AI tools and develop the talent needed to use them. But access means little when open opposition exists within academic culture.

The arts, humanities, and social sciences are no less important to Canada’s future than engineering, medicine, and technology. AI will reshape scholarship across every discipline. The challenge is that acceptance of AI is far less certain outside STEM.

Social scientists argue that AI distorts the conditions needed for meaningful research. They frame its use as incompatible with human inquiry and a threat to the integrity of qualitative scholarship. Refusing AI has become a statement about what it means to be a responsible researcher. 

These are not isolated cases. Over 400 researchers have signed a statement arguing that AI is inappropriate for all phases of reflexive inquiry. They reject AI on ethical grounds, arguing that its development is rooted in exploitative and colonial practices.

Pillar 4 sets out a necessary vision for AI-enabled research in Canada. But success will require more than sovereign infrastructure, talent development, and access to secure tools. It will require a research culture that sees AI as a legitimate tool for discovery rather than a threat to established ways of working.

Policy Action: Create Dedicated AI Research Funding 

AI will reshape scholarship across every discipline, but resistance is concentrated within the arts, humanities, and social sciences. If Canada hopes to build world-leading AI research capacity, it cannot afford for an entire segment of the research community to remain disconnected from these methods.

Governments should work with funding agencies like SSHRC to create dedicated streams for AI-enabled arts, humanities, and social science research. Existing competitions should remain intact, but new funding should reward projects that integrate AI methods and train graduate students in responsible use.

Funded projects should be required to document how AI was used, where its limits were encountered, and how researchers remained critically engaged with its outputs.

MAKING ROOM FOR CHANGE

Ottawa is investing in the infrastructure, talent, and tools needed for Canada to lead in AI. The next challenge is cultural. Universities must prepare students for an AI-enabled future and create research environments where new methods can be explored rather than dismissed.

The recommendations outlined here are intended to encourage that shift. For the strategy to succeed, universities must be willing to change with the technology.

AUTHORS’ NOTE

As social scientists who completed, or are completing, PhDs during the rapid emergence of AI, we approached these tools with curiosity. We tested where they helped, where they failed, and how they could be used responsibly, knowing that even exploring them could carry reputational risks.

Throughout our work, we encountered researchers arguing that AI has no place in qualitative inquiry, effectively calling into question the value of our own methods and findings. When one article made that case in absolute terms, we proposed a rebuttal grounded in our experience. It was rejected on the grounds that the debate had reached an acceptable conclusion, one that treated AI as an illegitimate research tool.

Such certainty is easier to maintain from the security of an established academic career. For students and early-career researchers, however, discouraging AI literacy may do more harm than good, regardless of discipline. As more graduates build careers beyond academia, AI skills are increasingly becoming part of the modern workplace. We have seen this expectation firsthand in our own job searches.