Universities are deploying generative AI in classrooms, admissions, and research, but most lack the governance to control it. A new framework presented at the AI in Higher Education Summit in Paris argues that institutions must balance three obligations: accountability, bold deployment, and creative redesign.
AI is already here—governance isn’t
Louis-David Benyayer, a big data and AI researcher at ESCP Business School, organized the summit in March 2026. The gathering showed that while many universities have AI strategies, few have policies to enforce them. The focus has moved from adoption to control and boundaries.
As the parallel sessions chair, I built the ABC Framework to address this challenge. The model requires institutions to govern AI with clear accountability, use it confidently, and rethink teaching and research around its capabilities. Missing any of these elements weakens the entire system.
Related: Bold Strategy Positions Higher Education as Success Lever
Accountability serves as the foundation. It determines what AI can decide, who reviews those decisions, and how affected parties can challenge them. The European Union’s AI Act, OECD principles, and UNESCO’s ethics recommendations all stress this point, yet most universities have not turned these guidelines into enforceable rules.
Bold deployment without rules is a liability
Confident use means AI in active roles: tutoring assistants, automated grading, and research tools. Tecnológico de Monterrey’s TECgpt platform, for instance, supports 90,000 students and 13,000 faculty. Without accountability, such systems could create institutional risks.
Creative redesign, the third element, reimagines education for an AI-integrated world. Summit discussions covered AI literacy, changes in teaching methods, and faculty training. Without governance, these efforts may lack consistency.
The framework’s key point is that the three elements must progress together. Accountability without use becomes hollow policy. Use without accountability invites harm. Redesign without either remains theoretical.
Related: Britain awakens to new European possibilities after Brexit
At the summit, papers on deployment described AI operating at scale, while governance discussions focused on draft frameworks—maturity models, accreditation pathways, and aspirational principles. The trend was clear: universities are implementing AI faster than they are establishing rules for it.
The gap reflects a deeper issue. AI governance is not standard risk management. It defines what AI does within an institution and who is responsible. That demands specific decisions, not vague principles.
Seven high-stakes decisions no university can avoid
AI governance involves seven areas where the technology directly affects rights, work, and status:
Related: The Transformative Power of Early Childhood Education: Building a Brighter Future
- Faculty use of AI, including disclosure requirements in syllabi and graded work.
- Student use, with integrity standards for coursework and research.
- Procurement, requiring audit rights and privacy-by-design reviews.
- Admissions and student services, where AI screening decisions must allow appeals.
- Hiring and promotion, where committees must oversee and audit AI-driven rankings.
These decisions cannot be left to IT departments alone. The most important choices are pedagogical and ethical, not technical. A diverse roundtable of faculty, staff, and students—both technical and non-technical—must shape governance to reflect the institution’s values, not just its systems.
The ABC Framework offers one way to structure this work. Other approaches will emerge. The choice is not whether to govern AI but whether to act now or fall behind later.
AI is already changing higher education. Governing it, using it, and redesigning around it form a single task, not three separate ones.
