Last term, a lecturer introduced an AI‑powered essay grading tool in our programme. She promised it would save time and provide instant, consistent feedback. At first, I was intrigued. I uploaded an essay, watched the screen blink for a few seconds, and then received a detailed report with a score, comments on structure, and even suggestions for improvement. But something felt off. My essay was 1,200 words long, clear enough, but nothing special. Yet the AI gave me a high mark and praised my “sophisticated vocabulary.” Then I noticed something stranger: when I pasted in a shorter, sharper essay from a friend—one that I knew was far better—the AI gave it a lower score, mainly complaining that it was “too brief.” I began to suspect that the tool wasn’t actually judging the quality of ideas, but was rewarding length, complex words, and a certain formulaic style. The more I tested it, the more it seemed like the algorithm was teaching us to write like a machine—not to think like a student.
That experience made me uneasy about the rush to bring AI into education. I started reading about EdTech—learning management systems, adaptive testing, AI tutors, digital assessment—and I realised that technology is not neutral. It embeds the assumptions of its designers, and when those assumptions are flawed, they get scaled across thousands of classrooms. I wanted to dig deeper, but I needed a focused, researchable topic. I began by browsing through collections of education technology (EdTech) research topics for 2026 (you can find them here: https://premierdissertations.com/education-technology-edtech-research-topics-2026/) to see how other students had turned similar concerns into proper academic projects. I found topics about algorithmic bias in grading, the effectiveness of AI feedback, the ethics of learning analytics, and the digital divide in access to educational tools. One topic caught my eye: “How does the use of AI essay graders influence students’ writing habits and their understanding of academic quality?” It felt exactly like the question that had been nagging at me. That spark gave me the confidence to design a small study around it.
Once I had my direction, I interviewed other students, collected examples of AI feedback, and compared it with the comments of human tutors. The patterns were clear. The AI rewarded long paragraphs, obscure vocabulary, and predictable structures, while human tutors valued clarity, originality, and genuine argument. Students who followed the AI’s advice started to sound more and more alike—polished but hollow. Some even admitted that they were writing to please the machine rather than to express their own thinking. My dissertation argued that while AI grading tools can be useful for basic feedback, they should never replace human judgement, and universities need to be transparent about how these tools are used. I recommended clearer guidelines for students, better training for staff, and a much more cautious approach to the adoption of EdTech in assessment. It was a modest project, but it grew directly from that first suspicious essay score.
Writing that dissertation changed how I see the technology in my classroom. I’m not anti‑EdTech—some tools genuinely help. But I’ve learned to ask hard questions about who designed them, what they reward, and who might be left behind. If you’re considering an EdTech research topic, start with a moment that made you uncomfortable: a platform that failed, a tool that seemed biased, a classroom where technology replaced conversation. The best research questions grow from that friction. Then browse real EdTech dissertation topics to shape your own inquiry. You don’t need to solve the future of education; you just need to ask one honest question, and then have the courage to follow it. Because technology should serve students, not the other way around.