A survey large enough to trust the regional breakdown
The Digital Education Council's 2026 global survey drew 45,398 total responses, split between 27,284 students and 18,114 faculty members, spanning 35 countries, a scale that gives its regional comparisons real statistical weight rather than the thin sample sizes that undermine many higher education AI surveys. That scale matters specifically for the survey's most striking finding: faculty intent to use AI declining in a specific region even as global and student-side adoption continues climbing.
The headline student number, 88 percent of students now using AI in their learning, represents a 16 percentage point jump from 2025, a rate of change that has clearly outpaced most institutions' ability to build corresponding guidance and policy infrastructure. The Digital Education Council's own framing captures that gap directly: AI has moved into the mainstream of student and faculty life faster than institutions have been able to respond to it, adoption is now widespread, but coherent practice is not.
The North American faculty pullback is the outlier finding
While overall faculty AI use in teaching sits at 77 percent globally, still below the 88 percent student figure, the more concerning data point is directional rather than absolute: faculty intent to use AI fell nine points specifically in the US and Canada, from 76 percent down to 67 percent, the lowest regional figure recorded anywhere in the survey. A declining intent figure in a region that started from a relatively high baseline suggests something specific happened in North American higher education over the past year to erode faculty confidence or willingness, rather than simply a slower initial adoption curve other regions are still working through.
That regional divergence is worth investigating directly for any US or Canadian institution's academic technology leadership. A global trend of rising faculty AI adoption running directly counter to a specific regional decline points toward local factors, policy uncertainty, academic integrity concerns, labor and job security anxiety among faculty, or accumulated frustration with unreliable early AI tools, that other regions in the survey may not share to the same degree.
The guidance gap sitting underneath both numbers
Two figures from the survey point to the same underlying structural problem from different angles: only 29 percent of students believe their instructors can actually guide them on appropriate AI use, and just 31 percent of faculty feel meaningfully involved in shaping their own institution's AI policy. Read together, those numbers describe a system where neither the people using AI daily, students, nor the people responsible for teaching its appropriate use, faculty, feel adequately supported by the institutional structures meant to govern that use.
For higher education technology and academic leadership, that dual gap argues against treating student-facing AI guidance and faculty AI policy involvement as separate initiatives. A faculty member who does not feel involved in shaping institutional AI policy is structurally unlikely to be equipped to confidently guide students on appropriate use, regardless of how much AI literacy training that faculty member individually receives, since policy involvement and practical guidance confidence are closely linked rather than independent capabilities.
The skills erosion concern cuts both directions
The survey captured a genuine, if asymmetric, concern about AI dependence affecting learning outcomes: 73 percent of faculty worry students are developing AI-dependent habits rather than building core skills, while a smaller but still meaningful 19 percent of students self-report retaining less information specifically because of their AI reliance. That gap between faculty concern and student self-reported impact is itself informative, suggesting either faculty are somewhat overestimating the skills erosion risk relative to what students themselves experience, or students are underestimating it in ways that will only become visible through downstream academic performance data.
Either interpretation argues for the same practical institutional response: building direct measurement of AI's actual impact on skill retention and academic performance, rather than relying on either faculty perception or student self-report alone to guide policy. Institutions that have not yet built that kind of direct measurement capability are currently making AI policy decisions based on two subjective, and apparently divergent, perception datasets rather than on outcome data.
What institutions should prioritize based on this data
For university and college academic technology leaders, this survey's most actionable finding is the specific North American faculty intent decline, since it represents a reversible trend if institutions can identify and address its specific local causes before it hardens into permanent faculty disengagement from AI-assisted teaching. The 88 percent student adoption figure is not going to reverse regardless of institutional response, which makes closing the guidance and policy involvement gaps the more urgent and higher-leverage intervention than trying to slow student AI adoption itself.
The practical starting point the data points toward is direct: institutions serious about closing this gap need to formally involve a meaningfully larger share of faculty in AI policy development beyond the current 31 percent, and need to invest specifically in faculty AI guidance capability rather than assuming faculty who use AI themselves are automatically equipped to guide student use, since the survey data suggests those two capabilities are not the same thing.
