Professors See Promise and Peril as AI Enters the University Classroom

Artificial intelligence is arriving in university classrooms faster than the rules that are supposed to govern it, and a new qualitative study from the United Arab Emirates offers a detailed portrait of how the people on the front lines feel about that mismatch. Researchers at Zayed University in Abu Dhabi interviewed faculty members and senior administrative staff in the university’s College of Technological Innovation to understand how educators perceive the opportunities and hazards of bringing AI tools into teaching. Their conclusion, published as an open-access research article in the journal Discover Education, is that educators see genuine promise in AI for personalized learning and workload relief, but they are held back by a tangle of unresolved questions about ethics, assessment integrity, data privacy, and who is actually responsible for making the technology work safely.

The study is distinctive for the theoretical lens it applies. Rather than simply cataloguing attitudes, the researcher used Cultural-Historical Activity Theory, a framework originally developed by Lev Vygotsky and later expanded by Yrjö Engeström, which treats human activity as a system of interconnected components: the people involved, the tools they use, the rules that constrain them, the community around them, the division of labor among them, and the outcomes they pursue. Within this framework, the most revealing findings are the contradictions, the tensions that arise when one component of the system falls out of step with another. In this case, the analysis identified several such tensions, including gaps between the AI tools now available and the institutional rules governing their use, and between educators’ enthusiasm for the technology and the absence of clearly assigned roles for training, policy, and support.

The empirical core of the study is deliberately small and deep. Five semi-structured interviews were conducted in person over five days in late March 2025, each lasting roughly thirty minutes, with three faculty members and two senior administrative staff selected for their direct experience with AI-related teaching and governance. The researcher analyzed the transcripts using reflexive thematic analysis, first coding line by line and allowing themes to emerge from the data, then mapping those themes onto the components of the activity system. By the fifth interview, no substantially new codes or themes were emerging, which the author cites as evidence that the small, information-rich sample was adequate for the study’s narrow aims. Ethical approval was obtained from the Zayed University Research Ethics Committee in February 2025, and the analysis was supported by reflexive memo-writing and peer debriefing with an AI specialist to guard against overfitting the data to prior expectations.

What emerged first was a picture of cautious optimism. Faculty participants described AI as naturally aligned with the digital habits of today’s students. One participant observed that the new generation of students lives in the cyber world and relies on it, suggesting that AI fits into an existing learning culture rather than imposing an alien one. Another highlighted accessibility applications, noting that AI could support students with disabilities, for example by translating sign language or Braille. These perceptions map onto the community component of the activity system, where student characteristics and expectations shape how educators judge whether a technology belongs in the classroom.

Participants also framed AI as a pedagogical support tool rather than a replacement for teaching. One faculty member compared AI to the transformative arrival of the internet, describing it as a co-pilot that enhances educator productivity and helps students build the digital skills their future employers will expect. The same participant argued that universities have a duty to prepare graduates for workplaces where AI is increasingly embedded in professional practice. A related theme concerned the division of labor in teaching: several participants said AI could absorb repetitive tasks such as grading and administrative work, freeing instructors to spend more time working directly with students on their understanding of the material. Crucially, they insisted that this redistribution does not eliminate the human element of teaching, which they saw as irreplaceable in guiding and motivating learners.

The concerns, however, were just as prominent and, in the analysis, more structurally significant. Participants repeatedly pointed to the absence of clear institutional rules as the central obstacle to responsible AI use. One faculty member worried that students might use AI to cheat on assignments if clear boundaries were not established. Another said the university needed explicit guidelines about when and how AI can be used, admitting uncertainty about how far educators are supposed to go. A third flagged data privacy as a serious issue whenever students are asked to use AI tools, stressing that student information must be protected. These worries, shared across multiple participants, correspond to the rules component of the activity system and illustrate a secondary contradiction between tools and rules: the technology’s pedagogical possibilities have outrun the frameworks meant to govern them.

The study’s discussion section interprets these tensions as systemic rather than isolated. Beyond the tools-versus-rules contradiction, the analysis identified a tension between educators and the division of labor, in which faculty recognize AI’s potential but remain hesitant because responsibilities for training, policy development, and technical support are not clearly assigned among faculty, IT staff, and administrators. A further contradiction emerged between the community and the object of the activity: students often show enthusiasm for AI and sometimes rely on it heavily, which faculty fear could disengage them from classroom experience and erode critical thinking skills such as analysis, evaluation, and synthesis. A final tension between tools and community reflected uneven access to training and institutional support, which participants said produced inconsistent AI adoption across courses and, in turn, uneven student experiences.

These findings arrive against a backdrop of a rapidly growing literature on AI in higher education. The study’s review of research published from 2020 onward describes work on personalized and adaptive learning systems, intelligent tutoring systems, generative AI in flipped classrooms, and AI-powered tools for measuring student engagement in online settings. That literature generally reports positive associations between AI-supported techniques and student engagement, but it also flags persistent challenges: teacher readiness and limited experience with the tools, ethical concerns including plagiarism and bias, the digital divide, and open questions about long-term effects on learning. The Zayed University study adds a socio-cultural dimension to this picture, noting that the UAE context carries its own dynamics, including culturally driven patterns of student motivation, that could complicate technology integration in ways generic frameworks might miss.

The author is careful about what the study can and cannot show. Because it captured the perceptions of five participants in a single college at one institution, the findings are presented as context-specific patterns rather than generalizable conclusions, and the interviews’ thirty-minute length may have limited how deeply some structural contradictions could be explored. Most importantly, the study examined faculty and administrator perspectives only; it did not directly measure student engagement or learning outcomes, so none of the reported perceptions should be read as evidence that AI improves or harms learning. The author suggests that future research include larger and more diverse samples, longer or repeated interviews, student perspectives, and multiple institutions to test whether the identified themes and contradictions recur elsewhere.

Even with those limits, the study’s practical message is coherent and likely to resonate far beyond Abu Dhabi. The participants’ accounts converge on a set of institutional conditions they view as necessary for responsible AI adoption: clear regulatory frameworks that specify acceptable use and protect assessment integrity and data privacy, comprehensive professional development so educators can use the tools creatively rather than fearfully, redesigned assessments that measure collaboration, problem-solving, and responsible AI use rather than content memorization alone, and equitable access so that uneven resources do not widen existing educational disparities. The contradictions identified through the activity-theory lens are not framed as reasons to reject AI, but as dynamic sources of change, in keeping with the theory’s own premise that tensions within a system can drive its development. In that reading, the current friction between fast-moving AI tools and slow-moving institutional rules is not a failure of integration but the very mechanism through which universities will work out what meaningful, ethical, and pedagogically grounded AI adoption looks like.

Subject of Research: Faculty and administrator perceptions of AI tool integration in higher education analyzed through Cultural-Historical Activity Theory

Article Title: Faculty and administrative perspectives on artificial intelligence tool integration in higher education using cultural historical activity theory

Article References: Dawood, M. N. (2026). Faculty and administrative perspectives on artificial intelligence tool integration in higher education using cultural historical activity theory. Discover Education, 5(1), Article 1217. https://doi.org/10.1007/s44217-026-02356-2

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02356-2

Keywords: artificial intelligence, higher education, Cultural-Historical Activity Theory, Zayed University, faculty perceptions, academic integrity, assessment, data privacy, qualitative research, student engagement, educational technology, institutional policy