Lebanese Teachers Split Into Three Camps on AI in the Classroom

Artificial intelligence has swept into classrooms around the world with the force of a technological tide, but a new study from Lebanon reveals that the teachers standing at the front of those classrooms are not being carried along uniformly. Instead, they are splitting into three sharply distinct camps — and the divide says as much about institutions as it does about technology. The research, published in Discover Education, offers one of the first in-depth qualitative portraits of how schoolteachers in a crisis-affected, low-resource education system actually think about AI, and the picture it paints is both encouraging and sobering.

The study, conducted by Mireille Riachi of the University of Balamand, surveyed 28 in-service Lebanese schoolteachers drawn from public and private schools across the country. Every participant had at least five years of classroom experience, and the sample was deliberately varied across school sector, educational level, subject specialization, and region. Data came from an open-ended online questionnaire structured around four dimensions: teachers’ understanding of AI, their actual uses of it, their professional development needs and institutional barriers, and their ethical concerns about data privacy, algorithmic bias, and the preservation of human-centered teaching. Responses were analyzed through thematic analysis, with a subset of coded data independently reviewed to check consistency.

The headline numbers reveal a striking paradox. Eighty-two percent of the teachers reported being familiar with AI — they had read about it, encountered chatbots, or used tools labeled as AI-driven. Yet only 29 percent had actually integrated AI into their instructional practices, and most of those confined it to administrative chores: automated grading, timetabling, generating standardized feedback. Just 11 percent had ever received any formal training related to AI in education. In other words, awareness has raced far ahead of capability, leaving a wide gap between what teachers know exists and what they can competently deploy in service of learning.

That gap, the study argues, is not merely a matter of individual skill deficits. Drawing on the Technological Pedagogical Content Knowledge framework — the influential model holding that effective technology integration emerges from the interplay of pedagogical, content, and technological knowledge rather than technical proficiency alone — Riachi interprets the pattern as a structural misalignment. Teachers are being exposed to AI tools without the pedagogical scaffolding needed to align those tools with instructional goals. The result is a kind of low-risk, peripheral engagement: AI handles the paperwork while the actual craft of teaching remains untouched by it.

The institutional barriers teachers described were concrete and, in the Lebanese context, severe. Public school teachers in particular pointed to poor internet connectivity, insufficient classroom devices, and the absence of technical support staff. One participant captured the frustration bluntly: “We are told to innovate, but no one tells us how or gives us tools.” Ninety-three percent of respondents identified AI-specific training as a pressing need, and many insisted that such training must be anchored in pedagogical frameworks rather than reduced to software tutorials. The study situates these complaints within Lebanon’s broader digital education landscape, where national strategies exist on paper — including a National Artificial Intelligence Strategy and large-scale teacher training initiatives that reached over 16,400 educators between 2020 and 2021 — but where implementation remains fragmented and unevenly distributed between well-resourced private institutions and strained public ones.

From the narratives, three distinct orientations emerged. The first, labeled the Hesitant orientation, describes teachers with high awareness but low engagement: they acknowledge AI’s relevance but refrain from using it pedagogically, deterred by lack of training, fear of error, and anxiety about undermining student learning. “I keep hearing about AI, but I’ve never had a real session or training that shows me how it works in class,” one wrote. Another confessed, “It’s like we’re expected to perform magic without tools.” These teachers often reported feeling left behind by institutional digital transitions, their digital literacy confined to routine administrative tasks.

The second orientation, the Adopters, stand at the cutting edge of classroom experimentation. Often self-taught, they deploy tools like ChatGPT, adaptive quiz platforms, and speech-to-text software to personalize learning, support language development, and enrich formative assessment. “I use AI to tailor questions based on student levels — it saves time and improves engagement,” explained a French teacher at a private school. Another reported that ChatGPT helped a shy student write his first full essay. These educators embody what researchers call the “teacher as designer” paradigm, actively reinterpreting technological possibilities to serve pedagogical ends. Yet even they operate without institutional recognition or structured frameworks, relying on personal initiative in an environment that offers validation to almost no one.

The third orientation may be the most provocative: the Skeptics, who resist AI on philosophical, ethical, and pedagogical grounds. Their objections are not technophobia but a considered normative vision of what education is for. “What happens when a child is wrongly flagged by AI? Do we blindly trust the machine?” one asked. “I teach literature. No machine can replace the human connection that fosters discussion and critical thought,” insisted another. A Beirut public school teacher voiced a data privacy anxiety that echoes worldwide: “I don’t know where the data goes, and that worries me. I wouldn’t want student information used by third parties.” Others worried about the dehumanization of teaching itself — “AI may be smart, but it’s not humane. Students need empathy, not algorithms.”

Crucially, the study refuses to read these ethical objections as mere resistance to progress. Instead, Riachi frames them as evidence of professional reflexivity — teachers acting as ethical mediators rather than passive consumers of technology. The skeptics’ concerns about algorithmic bias, surveillance, and student autonomy align closely with international guidance from UNESCO and the OECD, both of which advocate human-centered approaches to AI in education. The study suggests that skeptics should be treated as ethical interlocutors, deliberately included in professional learning communities and policy discussions, because a technology rollout that silences its most thoughtful critics is a rollout primed to repeat their predicted mistakes.

The recommendations that follow are pointed. Professional development must move beyond tool-based training toward pedagogical design, critical digital literacy, and ethical reasoning — with differentiated pathways for each orientation: scaffolded entry points and mentoring for the hesitant, collaborative co-design spaces for adopters, and seats at the ethical policy table for skeptics. Institutions must supply the infrastructure, leadership, and curricular flexibility that make innovation a collective, supported process rather than an individual burden. And robust ethical frameworks covering data privacy, algorithmic bias, accountability, and student autonomy must be formulated and disseminated before, not after, AI becomes embedded in schooling. The deeper message travels well beyond Lebanon: sustainable AI integration in education depends on repositioning teachers not as subordinate technicians but as co-designers of learning, whose professional judgment — not the technology itself — determines whether intelligent machines strengthen or erode the fundamentally human enterprise of teaching.

Subject of Research: Teachers' orientations toward artificial intelligence integration in Lebanese schools

Article Title: A qualitative typology of teachers’ orientations toward artificial intelligence integration in Lebanese schools

Article References: Riachi, M. (2026). A qualitative typology of teachers’ orientations toward artificial intelligence integration in Lebanese schools. Discover Education, 5(1), Article 1185. https://doi.org/10.1007/s44217-026-02264-5

Image Credits: AI Generated

DOI: 10.1007/s44217-026-02264-5

Keywords: artificial intelligence, education, Lebanon, teachers, qualitative research, TPACK framework, professional development, data privacy, algorithmic bias, educational technology, Discover Education, human-centered AI