American Students Are Relying On ChatGPT - At Their Own Risk

Newsweek

6 hours ago

Based on facts, either observed and verified firsthand by the reporter, or reported and verified from knowledgeable sources.
Newsweek AI is in beta. Translations may contain inaccuracies—please refer to the original content.
The use of generative artificial intelligence by students has increased over the last two years, but now research has revealed what is driving the trend.
A new survey found that students appreciated the ability of large language models (LLMs) like ChatGPT to provide information without any judgement, with many respondents describing it as a "safe and supportive" learning tool.
Why It Matters
The use of artificial intelligence in academic work is one of the biggest ethical issues facing the education sector. Tools like ChatGPT, which are being updated regularly to be more intelligent, can serve a purpose in helping students' work, but there is a worry that an overreliance could lead to problems.
What To Know
Last year, a study in the journal Computers and Education: Artificial Intelligence outlined that of 490 university students, one in four respondents (23.1 percent) relied on ChatGPT for drafting assignments and writing homework.
That research has now been backed up by a new report in the Tech Trends journal, published in June this year, which found that 78.7 percent of respondents were using generative AI regularly for their studies.
"Particularly noteworthy is that students perceived GenAI as useful because they are not judged by it and because of its anonymity," the report read.
"Students generally feel comfortable using GenAI for either general or learning purposes, perceiving these tools as beneficial especially with regard to their anonymity and non-judgmental nature."
Photo-illustration by Newsweek/Getty/Canva
However, the reliance on AI can be a double-edged sword. Another study from MIT found that extended use of LLMs for research and writing could have long-term behavioral effects, such as lower brain engagement and laziness.
The study, released this week without peer review, indicated that an overreliance on tools like ChatGPT "could actually harm learning, especially for younger users."
It compared brain activity between students using ChatGPT and students using traditional writing methods. The study found that the AI-assisted writers were engaging their deep memory processes far less than the control groups, and that their information recall skills were worse after producing work with ChatGPT.
What People Are Saying
Akli Adjaoute, an artificial intelligence security expert and author of Inside AI, told Newsweek of another pitfall for students. He says generative AI remained influenced by human hands in its programming, and "cannot be trained to be completely free of bias."
He added, "This is not a bug, it just reflects our world. AI does not invent knowledge. It learns from data created by people. And people, even with the best intentions, carry assumptions, disagreements, and historical baggage.
"AI systems are trained on information from many sources: books, websites, job applications, police records, medical histories, and social media. All of this information reflects human choices, including what we believe, what we value, and who has held power.
"If the data contains stereotypes or discrimination, the AI will absorb it. In many cases, it does not just copy the bias; it amplifies it."
What Happens Next
ChatGPT and other LLM tools continue to be updated regularly, but the academic sector is not moving as fast, and there is still no united approach on how AI tools should be handled.