Ninety-four percent of higher education professionals in EDUCAUSE’s latest survey say they’ve used AI tools for work in the past six months. That number is striking on its own. What makes it genuinely alarming is the next one: only 54% are aware of any policies or guidelines meant to govern that use.
EDUCAUSE’s 2026 report, authored by Jenay Robert and produced in partnership with AIR, NACUBO, and CUPA-HR, surveyed 1,960 higher education professionals between September and October 2025. The picture it paints is one of an industry that has already adopted AI at scale, without building the institutional infrastructure to manage it.
I think this is one of the most useful pieces of data we’ve gotten on AI adoption in higher education to date, precisely because it doesn’t focus on students or classrooms. The sample skews heavily toward managers and directors (39%), professional staff (32%), and executive leaders (16%), with faculty representing just 12%. This is a report about the operational backbone of higher education, the people running offices, managing teams, and making procurement decisions. And those people are using AI constantly, without much structure around them.
AI Tools in Higher Education: Where the Action Actually Is
Robert reports that 73% of respondents who use AI tools do so daily or weekly. The most common uses are brainstorming (63%), drafting emails (62%), summarizing long documents or meetings (61%), proofreading (56%), and creating presentations (47%). Tasks like writing code (28%), analyzing quantitative data (30%), and creating learning activities (31%) fall much lower on the list.
The pattern is telling. AI in the higher education workplace has settled into communication and content management, drafting, polishing, condensing. These aren’t the high-impact applications institutions like to highlight in strategic plans. They’re the daily grind tasks that people have automated for themselves without much fanfare. One faculty-specific finding is worth noting: 63% of faculty reported using AI for creating learning activities or assessments, compared to just 32% of staff. The instructional use case is alive, but it’s concentrated among the 12% of the sample who teach.
The Policy Vacuum
The gap between adoption and policy awareness is, by the report’s own framing, urgent. Robert writes that “institutions should view the creation and socialization of work-related AI policies and guidelines as an urgent issue due to the risks associated with improper AI tools use.” And the data back that up.
When you disaggregate the 46% who are unaware of policies by role, the numbers get worse: 38% of executive leaders and 43% of managers don’t know about any work-related AI guidelines. These are the people most likely to hold decision-making authority over such policies.
Robert reads this as evidence that many institutions simply don’t have work-related AI policies, not that communication has failed. I find that conclusion credible, and it tracks with what I’ve seen in the policy literature. When I covered Luo’s (2024) review of 116 US university GenAI policies, the focus was overwhelmingly on student-facing academic integrity. Work-related AI use, the kind this EDUCAUSE report documents, barely appeared in those policies. Institutions built rules for the classroom and forgot about the rest of the campus.
Among institutions that do have policies, the orientation tends to be permissive (47%) or neutral (30%). Only 20% of respondents described their institution’s policies as restrictive. And even among those with policies in place, the effectiveness is unclear. Robert notes that “the efficacy of existing work-related AI policies and guidelines remains unclear,” with only about half of respondents feeling confident or very confident using AI tools under their institution’s current guidelines.
The Workforce Development Disconnect
The report finds that 92% of institutions have some kind of work-related AI strategy. The most common strategy elements are piloting tools (65%), evaluating risks and opportunities (60%), and encouraging staff to use AI (59%). Only 5% are discouraging or prohibiting AI use. So institutions are saying yes to AI at a strategic level. The question is what comes after the yes.
On workforce development, 69% of institutions say they’re upskilling existing staff as their primary approach. That sounds responsible until you look at how they’re doing it. The top method, selected by 80% of respondents, is encouraging faculty and staff to develop skills on their own.
In-house professional development comes second at 71%, but a substantial gap separates formal training from informal encouragement. I’ve written before about how AI fluency requires structured, intentional development (Yee et al./McKinsey, 2025), and the EDUCAUSE data confirm that most institutions haven’t committed to that level of investment. They’re telling employees to figure it out, and calling that a strategy.
The ROI picture makes the gap even clearer. Only 13% of institutions are measuring return on investment for AI tools. Robert calls this “a major area of opportunity for higher education leaders,” which is a polite way of saying that institutions are spending on AI without tracking what they’re getting back. Respondents in the report expressed concerns that enthusiasm is outpacing evaluation. One noted they’re still waiting for a use case that actually reduces costs.
The Independent Thinking Concern and Shadow AI
When asked about the most urgent risks of using AI for work, respondents identified increased misinformation (55%), use of data without consent (52%), and loss of fundamental skills requiring independent thought (51%). That third one caught my attention. A majority of the people using AI daily or weekly for their jobs are simultaneously worried that it’s eroding their capacity for independent thinking.
This tracks with the cognitive research I’ve covered on this blog. Fan et al. (2025) documented metacognitive laziness in students using ChatGPT, and Shaw and Nave (2026) introduced the concept of cognitive surrender to describe how AI gradually reshapes reasoning patterns. The EDUCAUSE survey suggests this isn’t just a student problem. The professionals themselves are feeling it.
And then there’s the shadow AI problem. 56% of respondents reported using AI tools that weren’t provided by their institutions. Robert flags this as a significant risk because those tools “may not have been evaluated on important metrics such as data privacy and cybersecurity, accuracy and reliability, accessibility, protection of copyright and intellectual property.”
Nearly a quarter of faculty (23%) said their institution doesn’t provide access to any of the AI tools they want to use for work, the highest percentage of any role. So institutions are in a bind: they’ve said yes to AI in principle but haven’t given people the specific tools they need, which pushes employees toward unvetted alternatives.
What This Means for AI Strategy in Higher Education
The EDUCAUSE report’s recommendations are practical: codify AI responsibilities in job descriptions, provide formal training, communicate policies clearly, measure ROI. I’d add one thing. Institutions need to stop treating work-related AI use as a separate policy domain from instructional AI use. The same questions about integrity, quality, and critical thinking apply when a department chair uses ChatGPT to draft a budget justification as when a student uses it to draft an essay. The cognitive risks don’t stop at the classroom door.
Eighty-six percent of respondents want to keep using AI tools. Only 11% are required to. The adoption is voluntary, personal, and largely unsupervised. Institutions can either build the structures to support that reality or continue pretending the strategic plan is enough.
References
- Fan, Y., Tang, L., Le, H., Shen, K., Tan, S., Zhao, Y., Shen, Y., Li, X., & Gašević, D. (2025). Beware of metacognitive laziness: Effects of generative artificial intelligence on learning motivation, processes, and performance. _British Journal of Educational Technology, 56 (2), 489–530. https://doi.org/10.1111/bjet.13544
- Luo, J. (2024). A critical review of GenAI policies in higher education assessment: A call to reconsider the “originality” of students’ work. _Assessment & Evaluation in Higher Education_, 49(5), 651-664. https://doi.org/10.1080/02602938.2024.2309963
- Robert, J. (2026). The impact of AI on work in higher education. Research report. Boulder, CO: EDUCAUSE. https://www.educause.edu/research/2026/the-impact-of-ai-on-work-in-higher-education
- Shaw, S. D., & Nave, G. (2026). Thinking fast, slow, and artificial: How AI is reshaping human reasoning and the rise of cognitive surrender. Working paper, The Wharton School, University of Pennsylvania. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6097646
- Yee, L., Madgavkar, A., Smit, S., Krivkovich, A., Chui, M., Ramirez, M. J., & Castresana, D. (2025, November). Agents, robots, and us: Skill partnerships in the age of AI.