During the AJPOR editorial board meeting last month, we discussed the use of ChatGPT and similar generative AI algorithms in articles submitted to AJPOR. We recognize that these tools can mask poor scholarship but can also help to make research more readable and available. We also know that generative AI will be available to all and that we need to decide how to best use it.

To test ChatGPT, I asked it to “Write a 200 - 250 word essay on the appropriate use of ChatGPT in academic research papers.” The text is below, along with my reactions to it. Take a look!

ChatGPT is a large language model developed by OpenAI that uses artificial intelligence to generate human-like text responses. While ChatGPT can be a useful tool for generating ideas and exploring topics, its use in academic research papers requires careful consideration.

I would say – very careful consideration. But, I wouldn’t say it can’t be used.

Firstly, ChatGPT should not be used as a primary source of information for academic research papers. While the model can provide insights and generate ideas, its responses are generated based on patterns in its training data and do not necessarily reflect accurate or reliable information.

This recommendation is correct. It’s likely that a paper based on ChatGPT would fail careful peer review because it would not include all relevant information. It’s also likely the article would seem similar to previous research and may appear to be plagiarism to a smart reviewer.

Secondly, ChatGPT can be useful for generating research questions and exploring potential avenues for research. Researchers can use the model to generate text on a topic and then evaluate the responses to identify potential gaps in the literature or areas where further research is needed.

Here I disagree with ChatGPT. Most researchers don’t need help generating research ideas. They have more research ideas than available time and resources. This point is not really relevant to using ChatGPT in articles submitted to AJPOR.

Finally, ChatGPT can be a helpful tool for improving the clarity and coherence of academic writing. Researchers can use the model to generate text and then edit and refine the responses to create more polished and effective writing.

This paragraph is important to consider. AJPOR receives many articles that need substantial English proofing before publication. Sometimes ideas translated from the first language to English are not clear. Researchers could potentially improve their papers by having ChatGPT revise their text to academic English.

In conclusion, while ChatGPT can be a valuable tool for academic research papers, its use should be carefully considered. Researchers should not rely solely on ChatGPT for information, but instead MAY [emphasis mine] use the model as a way to generate ideas and improve the quality of their writing. By using ChatGPT responsibly, researchers MAY improve their research and writing and contribute to the advancement of knowledge in their field.

The last paragraph has important points. One goal of AJPOR has been to make the work of Asian scholars more accessible across the world. Improvements in writing style could assist AJPOR in that effort.

I highlighted “MAY” in the paragraph because using ChatGPT or similar algorithms requires the researchers to carefully evaluate the text and determine if it’s correct and consistent with their research.

I am not yet fully convinced that it can work without causing other yet unknown problems but, in the long run, it’s likely to be another tool we will use to be more productive.

What do you think? Please send any comments to kennedyj@indiana.edu and I’ll summarize them in the next issue.