Come up with mechanisms to detect AI-driven work, says academic
Published on: Tuesday, February 14, 2023
By: FMT Reporters
Text Size:
There were concerns about the use of AI-driven ChatGPT in academic work.
PETALING JAYA:Universities must come up with a mechanism that can verify the authenticity of writings by students and researchers in an age where artificial intelligence (AI) can be used to produce such work, an academic said.
Zuriati Ahmad Zukarnain of Universiti Putra Malaysia said this in response to concerns that AI-driven ChatGPT can produce such reports.
“The originality (of academic) reports and papers is very important,” she said at a forum when asked about measures to prevent AI-aided plagiarism in academia.
“Therefore, there should be a mechanism introduced by the university (which) would figure out whether the works were done by the students from their research findings or through outsourced (assistance) and (AI) technology.”
ADVERTISEMENT
Zuriati added that these mechanisms could be implemented in classrooms as well as in non-academic activities. As for postgraduate students, their supervisors can use a particular mechanism to verify the originality of their work.
Launched last November, ChatGPT is designed to mimic human-like conversation. Its responses are almost indistinguishable from the human-written text.
Last month, a professor at University of Pennsylvania’s Wharton School found that ChatGPT was able to score between a B- and B for the school’s Master of Business Administration (MBA) programme.
At the same forum, data scientist and AI expert Poo Kuan Hoong said ChatGPT would be designed to be more intelligent in the next five years.
He said ChatGPT would be developed to be utilised in healthcare and other sectors.
ADVERTISEMENT
“This (AI) technology will be trained in more specific domains and can actually (provide) more reliable and accurate answers.”