Digital Education Council: 88% of Students Globally Use AI, but U.S. and Canadian Faculty Intent Is Falling

By Hector Herrera | September 14, 2026

88% of students worldwide now use AI in their academic work—but in the United States and Canada, faculty intent to expand AI adoption is declining, breaking from the global trend. A Digital Education Council survey spanning 35 countries documents that student AI use has reached near-saturation globally, while the educators who design assessments and set policy are pulling back in North America. The divergence is producing measurable signs of skills erosion and widening an assessment gap that institutions have yet to close.

Why This Survey Matters

The Digital Education Council is a global network of universities tracking how technology is reshaping academic practice. Its multi-country dataset—covering 35 nations—is one of the largest global snapshots of AI in higher education collected to date. Most prior research has been country-specific or limited to a handful of elite institutions. At 35 countries, the DEC data enables meaningful comparison between regions where policy is permissive and those where it is restrictive or absent.

The headline numbers: 88% of students and 77% of faculty globally now use AI tools in their educational work. Both figures exceed industry projections from two years ago. What the data makes clear is that the question of whether students will use AI is settled. The harder questions—who shapes how they use it, and to what end—remain largely unresolved.

Faculty Intent Is Declining in North America

The most consequential finding is not the global adoption numbers. It is that faculty intent to expand AI use is declining in the U.S. and Canada, even as the global average continues to rise.

Faculty in North America are not pulling back because of technology unfamiliarity. The DEC data points to three interconnected concerns driving the retreat:

Academic integrity under strain. When AI can produce a competent first draft in under a minute, traditional written assignments no longer reliably assess student capability. AI detection tools have documented false-positive rates and have come under scrutiny in university arbitration proceedings—creating legal and reputational risk for faculty who rely on them to enforce policies.

Early evidence of skills erosion. The survey identifies measurable decline in foundational writing and independent research skills among students who use AI heavily without structured guidance. The concern is not AI use itself—it is AI as a substitute for developing the underlying competency, rather than a tool that builds on it.

Assessment ambiguity. Faculty report no institutional consensus on what AI-assisted work is worth compared to independent work, and a lack of standardized rubrics for evaluating the two. Without shared frameworks, each instructor is making individual judgment calls—producing inconsistency within departments and across institutions.

The result: North American faculty are cautiously retreating at the precise moment students are accelerating.

The Skills Erosion Problem

The DEC findings on skills erosion deserve direct attention. They are preliminary—early signals rather than causal findings—but they track with what writing instructors and librarians have been reporting anecdotally since 2023.

The pattern appears to be concentration effects. Students who use AI heavily for drafting without engaging with source material or revision are the ones showing declining foundational skills. Students who use AI as a research companion—exploring ideas, stress-testing arguments, generating counterexamples—appear to develop competency at normal or faster rates.

The distinction is pedagogical, not technological. It depends on how AI use is framed and structured inside a course. That is precisely the role that declining faculty intent leaves unfilled.

The Assessment Gap Is Widening

The DEC survey documents a widening gap between the rate at which students are integrating AI tools and the rate at which assessment practices are adapting.

The gap creates predictable distortions across three student populations:

• Students using AI as a shortcut exit assessments with grades that do not reflect actual understanding
• Students using AI skillfully gain efficiency advantages that instructors cannot distinguish from stronger underlying capability
• Students avoiding AI entirely may be disadvantaged on assignments calibrated for AI-augmented output volumes

No existing assessment standard reliably distinguishes these groups. Most institutional AI policies in the U.S. and Canada were written in 2023 and have not been updated to reflect the capabilities of 2026 models.

What Institutions Should Do

Prohibition strategies have largely failed. AI is now embedded in the writing software, search tools, and productivity platforms students use across all coursework—not just assignments where it is explicitly invoked. Enforcement at scale is technically impractical.

The more effective response is a framework approach: define what AI literacy looks like for each discipline, embed AI competency into existing learning outcomes rather than treating it as an add-on, and build faculty professional development that directly addresses the assessment and integrity concerns driving North American retreat.

Several European and East Asian university systems have already moved in this direction, which may explain why faculty intent outside North America is holding steady or growing. The gap between these regions and the U.S. and Canada will widen if North American institutions do not respond before the 2027 academic year.

What to Watch

Fall 2026 is the first major test of whether North American institutions respond to DEC-scale evidence with structured AI literacy frameworks or continue patchwork, course-by-course policies. The DEC plans a second survey wave in early 2027, which will show whether the North American faculty decline is a temporary adjustment or a durable structural shift.

Watch for updated guidance from the American Association of Universities and the Association of American Colleges and Universities before spring semester. If those bodies issue only voluntary guidance rather than shared evaluation standards, the assessment gap will likely persist through 2027.

Sources: Digital Education Council survey via EdTech Innovation Hub