Faculty Are Pulling Back on AI. Students Aren't Waiting.
A 45,000-response global survey finds student AI use has hit 88%, while US and Canadian faculty intent to use AI fell nine points, mostly because they feel shut out of policy decisions.
The gap the headline numbers hide
On the surface, the Digital Education Council's AI in Higher Education Global Survey 2026 looks like a straightforward adoption story. Drawing on 45,398 responses from 27,284 students and 18,114 faculty across 35 countries, it found that 88 percent of students now use AI in their learning and 77 percent of faculty use it in their teaching, up 16 percentage points on 2025.
Look closer and the picture splits. In the US and Canada, faculty intent to keep using AI fell nine percentage points, from 76 percent in 2025 to 67 percent in 2026, the sharpest regional decline the Council has recorded. Students in the same institutions are not slowing down at all. The result, as the report's authors put it, is adoption without authority: AI has moved into the mainstream of student and faculty life faster than institutions have worked out how to manage it.
Where the retreat is sharpest
The regional split is stark. In Asia-Pacific, 57 percent of faculty say they are excited about AI and believe it can make learning more effective and accessible; in the US and Canada, only 26 percent say the same. Fifty-five percent of North American faculty believe AI poses a serious risk to human intellectual development, against 29 percent in APAC.
A separate survey of over a thousand US faculty, run by the American Association of Colleges and Universities with Elon University, found roughly 90 percent saying AI is weakening student learning. Taken together, the two pieces of research point the same way: a large share of North American academic staff have moved from cautious interest to open scepticism inside a single year.
Adoption is now widespread, but coherent practice is not.
Why faculty are stepping back, not forward
The Digital Education Council's data suggests the retreat is less about the technology itself and more about how little say staff have had over it. Just 31 percent of faculty agree their institution involves them meaningfully in shaping AI policy. Around 82 percent cite resistance or unfamiliarity among colleagues as a live barrier in their department. Meanwhile 57 percent of students report that the assessments they sit come with inadequate guidance on what AI use is actually permitted.
That combination is telling. Staff are not being asked what good AI-assisted teaching looks like, students are not being told what counts as acceptable AI use in coursework, and both groups are left guessing at rules set somewhere above their heads. Faculty who feel excluded from policy-making have less reason to invest in learning a tool their institution hasn't asked them to help govern, and every reason to default to suspicion when they see it in student work they cannot properly assess.
Why this matters beyond the lecture hall
This survey covers higher education, but the mechanism it describes is not unique to universities. UK schools have spent the past year building AI policies at pace, often from the top down, in response to safeguarding rules, exam board guidance and inspection frameworks. The Digital Education Council's finding is a warning about what happens when that process leaves the people actually using the tools out of the room: adoption keeps climbing among the group with least oversight (students) while the group meant to be exercising professional judgement (staff) quietly disengages.
A policy document that arrives without consultation does not stop staff encountering AI in marking, in student work, or in their own workload. It just means they encounter it without a shared sense of what is expected, which is precisely the condition this survey links to faculty scepticism hardening into retreat.
What to do
- Treat staff as co-authors of AI guidance, not recipients of it. The 31 percent figure is a warning sign for any institution that drafted its policy without a staff working group.
- Close the assessment guidance gap directly with students. If more than half are unclear what is permitted, the fix is specific, subject-level guidance, not a single school-wide statement.
- Separate scepticism about AI's effect on learning from resistance to using it professionally. Both are legitimate, but they call for different responses: the first needs evidence and debate, the second needs training and time.
What to watch
Watch whether the US and Canada figure is a blip or the start of a wider pullback. If faculty intent keeps falling while student use keeps rising, the gap between how AI is actually used in classrooms and how it is officially governed will keep widening, and that gap is where most of the current disputes over academic integrity and assessment validity are already sitting.
Sources: EdTech Innovation Hub, Digital Education Council survey, Forbes, Higher Ed AI Playbook