By Sunil Sharma
Head of Content & Editorial Integrity at CSMFL Publications
Generative AI has helped humans become capable of creating vast amounts of content within moments. This content can be text, images, artwork, or programming code. This generated content can be helpful for educational purposes as far as the same is backed by trusted sources and scientific evidence. But AI models can create content that, on the surface, seems to be genuine but is not factually correct; this is due to hallucination on the part of these AI models, which is a known issue being further worked on by AI scientists and trainers.
Even after ensuring that all created content is backed by scientific evidence, this AI-generated content still remains associated with many ethical issues. This is the issue of integrity and ethical behavior for authors submitting this AI-generated content for consideration for possible publication in scientific journals, academic books, and other scholarly publications.
The issue of integrity and ethical behavior is challengeable on the grounds of who is the real creator of this content. Can just prompting AI models to create text or other graphical content be considered as creation or even value addition at minimum?
While dealing with such AI-generated content, it’s not just the mechanical tone of the writing that is the result of AI generation, but the question of how much original value that piece of content carries. This question of original value can be answered by evaluating the content against the following two questions:
If the answer is no, there is no value addition. It’s just a random piece of content containing a topic which is already there in the public domain. Ethically, this isn’t fit for consideration for any scholarly publication.
Using AI tools for structuring your writings can always be helpful, and this can save you time and effort in having a well-structured base for writing your work. But what happens if you go all the way to prompting an AI model to get your work written by it in its entirety. It remains an ethical problem, and the authors are responsible for what they submit.
To bypass detection as AI-generated content, many authors turn to specific tools created for the purpose—the purpose of making AI-generated content sound like human-written content.
Again, this is an ethical issue on the part of those humanizing tools, but that’s not the point of discussion here. We’re more focused on what ethical issues there are on the part of authors submitting ‘AI-humanized-AI-generated’ content for publishing as journal articles, books, book chapters, and other scholarly publications.
AI humanizing tools use Natural Language Processing (NLP) algorithms to identify the mechanical patterns and phrases that are used by AI models to create content. Then these AI humanizing tools rephrase the content by using words and phrases more commonly used by humans. Human tone and emotional depth to sentences are also introduced by these AI humanizing tools to make the content sound like a human-written content. But the purpose is not generally perfected, as all humans have a unique tone and writing style, which is impractical to imitate with these AI humanizing tools.
In simple terms, it can be said that AI humanizing tools use the imperfections of the AI detection tools. But AI detection tools are evolving day by day, and with new upgrades, this humanized content is again detected as AI-generated. So, humanizing content with AI tools becomes a game of “cat and mouse” between AI detection tools and AI humanizing tools, where one player challenges the other player, who in turn upgrades their game style to beat the other.
So, what’s the solution? The solution is being and staying human. AI-humanized content has no match for human-written content.
At CSMFL Publications, using AI and AI-assisted tools and technologies solely for creating content is strictly prohibited. However, the authors can use AI tools and technologies for improving the readability and language of their original writings. Read more in our Policy on the use of generative AI tools and technologies in academic writing.
This article is part of writing resources brought to you by the Writing Resources Center of CSMFL Publications. For more resources, visit Writing Resources Center