Seven in 10 secondary school students have used large language models (LLMs) for their studies, according to a survey of more than 300 US pupils.
鈥淚 realised that a lot of the people around me were using large language models, and more specifically ChatGPT, for a lot of school assignments,鈥� says Tiffany Zhu, an 11th-grade student (equivalent to year 12 in the UK) at The Harker School in San Jose, California.
To learn more, she worked with academics at Carnegie Mellon University in Pittsburgh, Pennsylvania, and the University of California, Santa Barbara, to survey the prevalence of AI usage in the student body. At present, many US schools ban the use of generative AI tools like ChatGPT, although some are developing policies for responsible, limited usage.
Advertisement
They signed up 306 students in the 6th to 12th grades (equivalent to years 7 to 13 in the UK) from around the US to complete a survey about their AI use, for which participants were paid around $0.80 each.
71 per cent of the students said they had used LLMs at least once for their school work, while 9 per cent said they used them daily. Around 5 per cent said they paid for an AI tool. The children who were more frequent users of LLMs had tried testing their abilities on work for different subjects with mixed results, with some mentioning that the LLMs occasionally hallucinated in historical contexts or sometimes gave incorrect answers because of a lack of rigorous reasoning.
鈥淭he paper shows that high school students are actively using LLMs, with or without the knowledge of their schools and teachers,鈥� says at Imperial College London.
He says the education sector should think about equipping students with the knowledge to use AI properly. 鈥淭here鈥檚 quite a danger here where students are relying on LLMs without proper training and expecting fully formed answers, plus the potential for them to have gaps in their learning.鈥�
It is a concern that Zhu shares. 鈥淭here needs to be a way where students and teachers and AI developers are working collaboratively to understand how we can use AI,鈥� she says. 鈥淲e can tell from our results that students are already going to use them despite school rules or ethical consideration.鈥�
arXiv
