A major research funder in Australia has banned reviewers from using artificial intelligence chatbots after allegations that they were used to generate feedback.
Australian Research Council, after an applicant for a grant of up to A$500,000 (£262,000) awarded under the Discovery Projects scheme reported finding “obvious” signs of ChatGPT in the referee’s comments has issued new guidance on this issue.
of ARC Tracker Twitter Account The report found little evidence of criticism, insight, or appreciation, and was a “general backlash” of the application, with one reviewer calling it a “reproduction of responses” that appear at the bottom of all text produced on ChatGPT. ” said he even forgot to remove the prompt.
New guidance published on July 7 states that reviewers “must adhere to the principle of confidentiality.”
Campus Collection: AI Transformers like ChatGPT have arrived. What next?
“Releasing material to generative AI tools is a breach of confidentiality and reviewers must not use generative AI as part of their evaluation activities,” the policy states.
It further adds that judges are “required to provide detailed, high-quality and constructive evaluations to assist the Selection Advisory Committee in assessing the merits of the application.” The use of generation AI can undermine the integrity of the ARC peer review process by generating text with inappropriate content, for example, general comments and restatements of applications. ”
Under this policy, if the ARC suspects that a report was generated by an AI, the report will be excluded from the review process and the ARC “may impose consequential measures in addition to those imposed by the employing agency.” Yes.
Australian researchers have suggested that ChatGPT’s use to write feedback is a sign that academics in the country are under time pressure.
With regard to grant applicants, the ARC Guidance states that while AI “provides an opportunity to assist researchers in developing grant proposals,” it “could raise copyright and other intellectual property (including copyright) issues.” There is.” Content generated by Generative AI may be based on the intellectual property of others and may be factually inaccurate. ”
As such, the ARC “advises applicants to be cautious regarding the use of generative AI tools when preparing grant applications” and the university encourages all applicants to I have to prove that I am responsible,” he said. .
chris.havergal@timeshighereducation.com
