Key points are not available for this paper at this time.
Abstract Technological Advances in Artificial Intelligence (AI) have Brought forth the Potential for Models to Assist in Academic Writing. However, Concerns Regarding the Accuracy, Reliability, and Impact of AI in Academic Writing have been Raised. This Study Examined the Capabilities of GPT-4, a state-of-the-art AI Language Model, in Writing an American Psychological Association (APA)-style School Psychology Paper. Specifically, we Examined GPT-4’s Ability to Provide Citations and Substantiate Claims, Offer Factually Correct Information, and Generate Accurate APA (7th Edition; 2020) Style Citations and Referencing. The Results Reveal Limitations in GPT-4’s Performance, Including a lack of Proper Citations and Substantiation, Errors in Citations and Referencing, and the need for External Validation of Factual Accuracy. While GPT-4 Shows Potential, Further Refinement is Necessary to Enhance its Accuracy, Reliability, and Adherence to APA (2020) Guidelines. Future Research should Focus on Improving the Capabilities of AI Models for Scientific Writing and Developing Clear Policies for Their use in Academia.
Lockwood et al. (Sat,) studied this question.