Student Perceptions of Essay Grading with Artificial Intelligence in Higher Education

In the quiet hum of a university classroom, a familiar ritual is taking a new turn. For generations, the essay has been a bridge between a student's thoughts and a teacher's judgment, a process where a human reader weighs arguments, senses tone, and offers a personalized critique. This interaction is central to learning, yet it is also a heavy burden on educators, who must read thousands of pages to provide timely feedback. As artificial intelligence has matured, moving from simple calculators to systems capable of understanding complex language, researchers have begun to ask a practical question: can a machine take on the role of the essay grader? The answer is not a simple yes or no, but a complex landscape of hope and hesitation. To understand this, one must look at how people form their first impressions of new technology. Just as a traveler might judge a new city by its first street corner, students judge a new tool by their immediate experience with it. Their initial feelings—whether they feel the tool is fair, helpful, or cold—shape whether they will trust it in the future. This dynamic is crucial because if students do not believe the system is accurate or fair, they will not engage with it, no matter how efficient it is.

A team of researchers from universities in Austria and Finland set out to explore this exact moment of first contact. They did not simulate a classroom or ask students what they thought a machine might do. Instead, they placed 109 university students in a real learning situation. The students were asked to write a short essay on a topic concerning data-driven business opportunities. After they finished, they submitted their work to a commercial artificial intelligence system designed to grade essays. The machine immediately provided a numerical score and a written critique, along with a model answer showing what a high-quality response looked like. The students then answered a single, open question about their instant feelings regarding the experience. The researchers collected 109 written responses, ranging from a single word to several sentences, and used a sophisticated method to find patterns in what the students said. They did not simply count words; they grouped the responses by meaning, looking for the common themes that emerged when students faced a machine that claimed to judge their writing.

The analysis revealed seven distinct themes that captured the full spectrum of student reactions. On one hand, the students found the experience remarkably convenient. They appreciated the speed of the feedback and the ease of using the tool, noting that it felt user-friendly and saved them time. This positive reception suggests that the technology works well as a practical utility, removing the friction of waiting for a human teacher to return a graded paper. However, this convenience was shadowed by a deep and persistent unease regarding accuracy. A significant number of students expressed doubt about whether the machine could truly understand the nuances of their writing. They worried that the system might miss the subtle points of an argument or fail to recognize creativity, leading to a score that felt unfair. This skepticism was not just about the grade itself, but about the trustworthiness of the system; many students felt they were being judged by a black box that they could not fully understand or question.

Perhaps the most telling finding was the students' strong desire for a human touch. While they acknowledged the efficiency of the machine, many felt that the feedback lacked the empathy and depth of a human teacher. They sensed that a machine could not truly understand the context of their struggle or offer the kind of encouragement that fosters growth. This was not a rejection of the technology, but a clear boundary drawn around its role. The students seemed to view the AI as a useful supplement, a tool that could handle the initial pass or provide quick practice, but not as a replacement for the human element in education. They wanted the machine to do the heavy lifting of checking facts or structure, but they wanted a person to handle the judgment of quality and the emotional connection of learning. The study suggests that the future of grading lies not in choosing between humans and machines, but in a hybrid approach where both work together.

The researchers concluded that while the technology is ready to assist, it is not yet ready to stand alone in high-stakes situations. The students' mixed feelings highlight a critical gap: the technology is fast and convenient, but it has not yet earned the full trust of the learners it serves. The study does not claim that artificial intelligence is bad for education, nor does it suggest that human teachers are obsolete. Instead, it points to a need for balance. Educators must design systems where the machine provides immediate, consistent feedback while a human teacher oversees the process, ensuring fairness and adding the necessary empathy. The findings serve as a guide for how to introduce these tools without breaking the trust between student and teacher. By listening to these initial reactions, schools can build a future where technology supports learning without losing the human connection that makes education meaningful.