When I first heard about the school introducing an AI tutor to help struggling readers, I was fascinated. I imagined a patient, tireless machine that could adapt to each child’s pace, offering endless encouragement without ever losing its temper. As a student of education, I believed technology could be the great equaliser — a way to close the gap between the privileged and the underserved. That belief lasted exactly one afternoon. I was sitting in on a session where a quiet, anxious boy named Ethan was trying to read a short paragraph aloud to the AI. The software corrected him with a cheerful chime — “Almost! Try again!” — but Ethan didn’t smile. He didn’t try again. He put his head on the desk and refused to speak for the rest of the session. Later, his teacher, a woman in her sixties named Mrs. Jennings, crouched beside him and said nothing. She just placed her hand gently on his back and waited. After a minute, Ethan looked up. “I can’t do it,” he whispered. Mrs. Jennings said, “You already did four hard things today. Let’s try one more together.” And Ethan, slowly, picked up the book again. The robot had failed where a human had succeeded — not because the technology was bad, but because it couldn’t understand that Ethan needed compassion, not correction. That moment reshaped my entire view of educational technology, and when I had to choose a dissertation topic, I knew I wanted to explore something at the intersection of human connection and digital tools.
The more I thought about that afternoon, the more questions I had. Can AI tutors ever replicate the emotional intelligence of a skilled teacher? What do we lose when we replace human interaction with algorithmic feedback? And in an era of rapid technological change, how should teacher training programmes evolve to prepare educators for classrooms where AI is increasingly present? I didn’t want to either glorify technology or demonise it — I wanted to understand where it genuinely helped and where it fell short. To find a specific, researchable angle, I spent some time exploring what other students had already investigated. I came across a collection of education dissertation topics 2025 that gave me a clearer picture of the current landscape. Some projects examined the effectiveness of gamified learning platforms in primary maths, others explored teacher attitudes toward generative AI in assessment, and a few looked at how remote learning during the pandemic had changed parent‑teacher relationships. That breadth helped me narrow my focus to a question that felt both urgent and under‑researched: how do primary school teachers perceive the role of AI tutoring tools in supporting children with low self‑efficacy in reading?
I designed a small qualitative study, interviewing fifteen primary school teachers across the West Midlands who had used at least one AI‑based reading tool in their classrooms. The conversations were rich and often surprising. Many teachers described the AI as a “useful assistant” — good for phonics drills, for tracking progress, for providing data that helped them identify struggling students early. But almost every teacher drew a firm line. The AI could reinforce skills, they said, but it couldn’t build confidence. It couldn’t notice when a child was tired, or hungry, or on the verge of tears. It couldn’t forge the kind of trusting relationship that made a child willing to fail publicly and try again. One teacher put it bluntly: “The robot can teach them the sounds. But it can’t teach them that they matter.” That sentence became the emotional core of my dissertation. I argued that while AI tools have a legitimate place in the modern classroom, they should be deployed as supplements, not substitutes — and that teacher training programmes need to equip educators with the critical digital literacy to make informed decisions about when, and when not, to use them.
Writing the dissertation didn’t give me a simple answer to the big questions about technology and education. It gave me something better: a framework for asking the right questions, and a deep respect for the Mrs. Jenningses of the world who do what no algorithm can. If you’re considering an education dissertation in 2025, don’t be afraid to tackle the tensions that are shaping classrooms right now — between tradition and innovation, between human and machine, between the evidence‑based and the intuitive. The best research often emerges from moments of discomfort, when something you believed in is challenged by something you witnessed. Follow those moments. They’ll lead you somewhere worth going.