Illustration by Celine Bejjani
How do we measure the impact of artificial intelligence on our lives? Most studies begin with numbers: How many students use AI? How often? For what tasks?
Surveys and statistics can tell us how widespread these tools have become. But numbers alone cannot explain how people relate to AI, what it replaces or what it subtly reshapes in their daily habits. Measuring use is not the same as understanding experience.
This is where qualitative research — particularly ethnography — becomes essential. Ethnography involves close observation, in-depth interviews and long-term engagement with people in their everyday environments. Instead of asking only “how much?”, it asks “how?” and “why?”
Consider a common scenario. A student posts a question in a WhatsApp study group. Hours pass without a response. The student turns to ChatGPT: “Can you help me with this assignment?” The chatbot answers in detail. A few minutes later, the tone shifts: “ I’m stressed.” The reply comes back instantly, empathetic and structured.
A quantitative study might record two uses of AI: one academic, one emotional. An ethnographic study would probe further: Why turn to a machine instead of peers? Was it speed, privacy, embarrassment or convenience? How does repeated reliance alter peer networks or confidence over time?
Without interviews and close observation, we risk misunderstanding what AI is actually doing in students’ lives.
Constraints of numbers
Quantitative research — surveys, statistics, etc. — is useful. It can show trends, reveal scale and inform institutional planning. But numbers cannot easily capture motives, emotions or subtle social shifts. They cannot explain how students feel when they rely on AI, whether they experience relief, guilt or dependency. Nor can they illuminate what changes gradually, such as shifts in critical thinking habits or collaboration patterns.
Reports have identified a “blind spot” in assessing AI’s everyday effects. Existing approaches tend to focus on compliance and regulation, ensuring AI systems follow laws; on embedding ethical values into design, like privacy or fairness; or on philosophical thought experiments about morality in decision-making. While valuable, these methods often overlook the mundane reality of daily interaction: the quiet moments when a student chooses a chatbot over a classmate. It is in these ordinary decisions that larger social transformations begin.
Lebanon’s qualitative strength
In Lebanon, large-scale quantitative research infrastructure remains limited due to funding constraints and institutional fragility. As a result, many researchers rely heavily on qualitative methods, particularly ethnography. Over time, this reliance has cultivated methodological sensitivity to context, relationships and social nuances.
Through in-depth interviews, careful observation, and extended engagement, ethnographers uncover layers of meaning hidden from raw data.
For instance, observing students in libraries or informal study spaces may reveal patterns invisible to statistics: who asks AI for help, which assignments trigger stress, when emotional reassurance is sought and how these patterns differ across socio-economic or gender lines.
Ethnography captures these behaviors because it examines human-AI interactions as embedded in social life — it sees the student, the class, the WhatsApp group and the broader cultural context as intertwined with the technology itself. It also captures change over time: how repeated reliance on AI might affect critical thinking, collaboration or peer relationships. It can show how students gradually begin to depend emotionally on a machine — a phenomenon that statistics alone cannot capture.
What ethnography actually looks like
Ethnography is not a quick survey. It requires time, trust and attentiveness. Researchers may spend weeks or months observing, interviewing and reflecting on what they see. They seek to make connections between what people say and what they do, capturing nuances like hesitation, inside jokes, moments of silence, or even what is absent but felt — the “absent-present” data.
And the answers may not look the same everywhere. A student at the American University of Beirut might use AI to refine essays supported by strong English-language training and institutional tutoring systems. A student at the Lebanese University, navigating crowded classrooms and fewer university resources, may rely on it differently — perhaps as a substitute for limited faculty access, or as a shortcut when time and support are scarce.
A quantitative study might conclude that students at both universities “use AI frequently.” It will struggle to reveal how institutional culture, socio-economic background, language fluency or access to support shape that use.
Ethnography can, through observation, interviews and sustained engagement, illuminate how AI use is woven into broader inequalities and everyday realities within Lebanon’s higher education system.
Ethnographies can help inform policies
The Lebanese educational system is only beginning to grapple with AI. Schools and universities are exploring integration, yet comprehensive national guidelines remain limited. Ethnographic research can inform more grounded policy responses.
By observing and documenting students’ interactions, educators and policymakers can design frameworks that are more comprehensive and informed. To name a few, policies can encourage the use of AI as a tool, not a replacement for human learning; monitor emotional reliance to prevent unhealthy attachment; support peer collaboration and critical thinking; inform training for teachers to guide students effectively
Beyond the use
Globally and in Lebanon, AI is no longer a simple technological innovation — it has become a social phenomenon embedded in education, work and emotional life. Studying it through quantitative methods offers valuable but partial insight. Ethnography complements this by revealing hidden dimensions of human-AI interaction: the gradual displacement of peer networks, the subtle shift in cognitive habits and the socio-cultural effects of automation in education.
Ultimately, the question is not whether students use AI — they do. The deeper questions dig into what they gain, and what they quietly give up in return.
By combining quantitative measures with ethnographic insight, Lebanon can approach AI not merely as a technical challenge but as a social transformation requiring thoughtful guidance.
