Physics university students' perceptions of AI

In the quiet hum of a university library or the glow of a laptop screen late at night, a new kind of helper has arrived for students. This helper is artificial intelligence, a technology that can read, write, and solve problems much like a human, but at a speed and scale that feels almost instantaneous. For decades, universities have been places where young people learn to think for themselves, to struggle with difficult concepts until they click, and to prove their understanding through exams and essays. Now, that struggle has a new variable. The question facing educators and students alike is no longer whether this technology exists, but how it is changing the very nature of learning. When a student asks a computer to explain a complex idea or to check their work, are they learning faster, or are they skipping the steps that build a strong mind? This tension between efficiency and deep understanding sits at the heart of a recent investigation into how physics students are navigating this new landscape.

Researchers at King's College London set out to understand the reality of this shift among undergraduate physics students. They did not just ask if students were using these tools; they wanted to know why, how it felt to use them, and what the students feared for their future. The team combined a broad survey of seventy-one students with intimate, hour-long conversations in small groups. They listened to the students' own words, looking for the patterns that numbers alone could not reveal. What they found was a picture of a student body that has almost universally adopted this technology, driven by a mix of genuine need, social pressure, and a deep, often unspoken anxiety about the future.

The study revealed that nine out of ten physics students are using artificial intelligence regularly. For more than a third of them, it is a daily habit; for the rest, it is a frequent companion. These students are not using the technology merely to bypass work. Instead, they turn to it for two main reasons: to understand difficult concepts and to save time. Many students described using the tools to rephrase dense textbook language into something clearer or to get hints when they were stuck on a problem. They valued the ability to get immediate feedback without the fear of looking foolish in front of a professor or a peer. One student noted that the tool felt like a judgment-free space where they could ask "stupid questions" without the shame of admitting they didn't understand. In this way, the technology acted as a personal tutor that was always available, filling gaps that the traditional classroom sometimes left open.

However, this convenience came with a heavy cost that the students themselves were acutely aware of. A significant portion of the group felt that relying on these tools was eroding their own ability to think critically and solve problems on their own. They worried that by letting the computer do the heavy lifting, they were losing the mental muscle required for deep learning. This fear was not just about grades; it was about their identity as learners. Some students admitted that using the tools made them feel like impostors, as if their achievements were not truly their own. They also noted a strange social dynamic at play: even students who disliked the technology felt they had to use it. They feared that if they refused, they would fall behind their peers who were using the tools to work faster and smarter. It became a self-fulfilling prophecy where everyone used the technology simply because everyone else was using it.

The conversation in the focus groups moved quickly from the classroom to the wider world, where the students' anxieties about their careers took center stage. Many expressed a profound worry that the very jobs they were training for might not exist when they graduated. They spoke of a future where artificial intelligence could replace not just entry-level tasks, but entire fields of research and theory. One student recounted a conversation with a researcher who suggested that theoretical physics might be obsolete within a few years. This fear of displacement was not abstract; it was a daily worry that colored their motivation to learn. Alongside these career fears, the students also raised concerns about the environment. They were aware that the massive data centers powering these tools consume vast amounts of water and energy, a cost that seemed to contradict the clean, efficient image of the technology they were using.

When asked how universities should respond to this new reality, the students offered a clear and surprising solution. They were skeptical of the idea that schools should simply teach them how to use the tools better, arguing that the technology changes too fast for any single method to remain relevant. Instead, they pointed toward a return to older, more human forms of assessment. Many students expressed a strong preference for oral examinations, where a professor could ask them to explain their thinking in real time. They believed that while a computer could write an essay or solve a problem on a screen, it could not fake the depth of understanding required to hold a conversation about a complex topic. They saw these face-to-face interactions as the only way to prove that they truly knew the material, rather than just knowing how to prompt a machine.

The study concludes that the rise of artificial intelligence in education is not just a technical issue, but a social and psychological one. The students are not passive users; they are navigating a complex web of efficiency, fear, and peer pressure. They recognize the benefits of the tools but are deeply concerned about the long-term effects on their skills and their futures. The researchers suggest that universities must look beyond simple bans or permissions and consider how to protect the integrity of learning in an age where answers are instant. The students' voices indicate that the path forward may lie in assessments that value human conversation and critical thought, ensuring that the next generation of physicists remains capable of thinking for themselves, even in a world full of smart machines.