AI advance with ChatGPT warns of student cheating

Universities are being urged to guard against the use of artificial intelligence (AI) in newsrooms, following the emergence of a sophisticated chatbot program that can mimic academic work. This sparked a debate about the best ways to assess students in the future.

The ChatGPT program, created by the Microsoft-backed company OpenAI, can craft arguments and write compelling text, causing widespread concern that students will use the software to cheat on written assignments.

Academics, higher education consultants, and cognitive scientists around the world have suggested that universities develop new modes of assessment to counter the threat posed by AI to academic integrity.

ChatGPT is a large language model trained on millions of data points, including large blocks of text and books. It generates convincing, coherent answers to questions by predicting the next plausible word in a string of words, but its answers are often inaccurate and require fact-checking.

When the program is asked to produce a reading list on a particular subject, for example, it can generate false references.

This week, around 130 university representatives attended a seminar by JISC, a UK-based non-profit body that advises higher education in technology. They were told that a “war between plagiarism software and generative AI won’t help anyone” and that the technology can be used to enhance writing and creativity.

The wide accessibility of this tool, which is free to the public, has raised questions about whether it makes essays redundant or requires other resources to correct content.

Turnitin is software used by around 16,000 school systems around the world to detect plagiarized work and can identify some types of AI-assisted writing. The US-based company is developing a tool to guide educators in evaluating work with “tracemarks” of AI, said Annie Chechitelli, chief product officer at Turnitin.

Chechitelli also warned against an “arms race” in detecting cheaters and said educators should encourage human skills such as critical thinking and copyediting.

Overreliance on online tools can affect development or creativity. A 2020 study from Rutgers University suggested that students who use Google to answer their homework score lower on exams.

“Students won’t automatically get an A when presenting AI-generated content; it’s more of a ‘workhorse’ than an Einstein,” said Kay Firth-Butterfield, head of artificial intelligence at the World Economic Forum in Davos, adding that the technology will improve quickly.

Academics have warned that education has been slow to respond to these tools. “The education system in general is just waking up to this, [mas é] the same kind of problem as cell phones at school. The response was to ignore it, reject it, ban it, and then try to accommodate it,” said Mike Sharples, professor emeritus at the Open University and author of “Story Machines: How Computers Have Become Creative Writers” [Máquinas de histórias: como os computadores se tornaram escritores criativos]🇧🇷

Shifting to more interactive assessments or reflective work can be costly and challenging for an already cash-strapped sector, said Charles Knight, a higher education consultant.

“The reason essay writing is so successful is partly economic,” he added. “If you do [outras] assessments, the cost and time required increases.”

Universities UK, which represents the sector in the UK, said it was closely following but not actively working on the issue, while Australia’s independent higher education regulatory body, TEQSA, said institutions need to clearly define their rules and communicate them to the students.

“Learning is a process, it’s not the end result in many cases, and an essay is not useful in many jobs,” said Rebecca Mace, a digital philosopher and educational researcher at the Institute of Education at University College London (UCL).

Translated by Luiz Roberto M. Gonçalves