Can scholars write journal papers using AI? What the guidelines say

Artificial intelligence (AI) refers to “intelligent machines and algorithms that can reason and adapt based on a set of rules and environments that mimic human intelligence.” This field is evolving rapidly, with one education sector being Buzz in discussions about AI use for writing.

This is important for people who rely on reliable information, not just academics, but also from journalists and policymakers to educators and the general public. Ensuring transparency in how you use AI protects the reliability of all publicly available knowledge.

In education and research, AI can generate texts, improve writing styles, and even analyze data. Save time and resources by allowing quick summary of work, language editing, and reference checks. It also enhances academic work and can inspire new ideas.

Similarly, AI can generate the entire work. It can be difficult to distinguish between original works written by individuals and work generated by AI.

This is a serious concern in the academic world for universities, researchers, lecturers and students. Some uses of AI are considered acceptable, while others are not (or not yet).

In my ability as an editor and editorial board member of several journals, and as a psychology researcher and professor, I have worked on what is considered an acceptable use of AI in academic writing. I have turned to various published guidelines:

The guidelines are unanimous that AI tools cannot be listed as co-authors or held responsible for their content. The author is entirely responsible for verifying the accuracy, ethical use and integrity of all AI-influenced content. Regular support does not require citations, but you need to clearly refer to content generated by substantive AI.

Let’s unpack this a little more.

Assist vs. Generated Content

When understanding AI’s use in academic writing, it is important to distinguish between AI-assisted content and AI-generated content.

AI-assisted content refers to work primarily written by individuals, but has been improved with the help of AI tools. For example, authors may use AI to assist with grammar checks, to enhance sentence clarity, or to provide style suggestions. The author maintains control and AI simply acts as a tool to hone the final product.

This type of support is generally accepted by most publishers and publishing ethics committees without requiring formal disclosure. That is to the extent that the work remains original and the integrity of the research is supported.

AI-generated content is generated by the AI ​​itself. This means that the AI ​​tool generates a substantial portion, or entire section, based on detailed instructions (prompts) provided by the author.

This raises ethical concerns, particularly regarding originality, accuracy and authors. Generated AI draws content from a variety of sources, including web scraping, public datasets, code repository, and user-generated content. This is accessible content. I’m never sure about the reliability of my job. AI “Hazardization” is common. Generic AI may be plagiarizing other people’s work or infringing copyright, and you may not know.

Therefore, in the case of AI-generated content, the author must make clear and explicit disclosures. In many cases, this type of content can face limitations. Publishers may even reject it completely, as outlined in the Publication Ethics Guidelines Committee.

What is allowed, what is not

Based on the guidelines reading, we provide practical tips for using AI in academic writing. These are fairly simple and can be applied to fields.

All guidelines state that AI tools can be used for everyday tasks such as improving grammar, revising sentence structures, and assisting in literature searches. These applications do not require specific approval. Throughout the guidelines reviewed, this is necessary for research and AI-generated content is not permitted unless there is a reason for the content to be clearly marked and referenced. Therefore, depending on how you use AI, you may need to refer to it in your manuscript. This can be found in the literature review, or in the Methods or Results section. By citing this appropriately, Sage and the Publication Ethics Committee emphasize that authors must disclose when AI-generated content is used. There are different rules for citing AI use, but it seems all agree that you need to quote the name of the generator used, the access date, and the prompt used. This level of transparency is necessary to maintain the reliability of academic work. Other aspects linked to AI support, such as code modification, tables and diagram generation, reduced word counts, and checking analysis, cannot be directly referenced in the body of the manuscript. This should be presented at the end of the manuscript in line with current best practice recommendations. The author checks the accuracy of AI content, whether AI assisted or generated by AI, to ensure that there are no bias, plagiarism, and potential copyright infringement.

The last word (for now)

AI tools can undoubtedly enhance the academic writing process, but they must be used with transparency, caution and respect for ethical standards.

The author must remain vigilant about maintaining academic integrity, especially when AI is involved. Authors should verify the accuracy and appropriateness of AI-generated content and not compromise on the originality or validity of their work.

There were excellent suggestions on when AI declarations are required, optional, and not necessary. In cases of uncertainty, the best advice is to include the use of any form of AI (support or generation) in your approval.

As AI continues to evolve, it is highly likely that these recommendations will be revised in time. But it’s just as important to start somewhere. AI tools stay here. Let’s deal with it constructively and collaboratively.

This article will be republished from the conversation under a Creative Commons license. Please read the original article.