AI for academic writing: friend or foe?

A biased friend.

I think generative AI can be a very useful tool to begin a project (grant application, manuscript preparation, etc.) in terms of: ensuring there are not important/relevant existing data of which the investigator was not aware, organizing initial/general ideas into a proper scientific format, and beginning to link thoughts together.  Given the ubiquitous examples of absurd errors with answers to certain questions, though, it is clear that this helpful tool has not yet reached a standalone status and MUST have careful human review and likely substantial editing before submission, both to ensure scientific accuracy and also logical connections.  That being said, it always feels easier and more efficient (to me, at least) to edit a draft than to create something new on a blank page. The activation energy to overcome that lone blinking cursor sometimes feels very high, and I think of this in a similar way to a trainee creating a draft manuscript or grant application that needs mentor review, sometimes in a substantial manner.

Then, if the human expert-edited document is fed back into the system, and this is done enough times within a particular AI program, the program will begin to "learn" what previous associations are not truly valid, etc.  At least that's my understanding of how it works...

I see others mentioning AI (specifically ChatGPT) as a good tool in some circumstances, in particular for researchers whose first language isn't English. Personally, I'm very skeptical of its value in that regard - I have yet to see any output from ChatGPT that reads well. I don't think anyone's expecting it to produce literary masterpieces but not even simple tools like Grammarly have a perfect grasp of English, and ChatGPT can't seem to produce a high-school essay level of writing quality, let alone content. Is that really a good tool for non-English speakers? Isn't that just going to disadvantage them further in a scientific field full of native English speakers? Shouldn't we (the humans) be making scientific writing more accessible anyway?

Personally, I've only found AI useful in one very limited context: to me get past the "staring at a blank page" stage of writing. I've tried to have it summarise texts for me, but its summaries are much worse than the abstracts written by humans. I've found tools like ResearchRabbit (sort of the AI version of ConnectedPapers) useful, but that's really just sprinkling AI onto a task that standard computing was already much better suited for anyway - trawling thousands of data points (papers) and making simple connections based on shared authors or citations. 

If I had to pick, I would call AI a "friend" to researchers, but not a particularly useful one. That might change in the future, but I don't really see it breaking machines out of their well-established niches within science anytime soon.

Is good. but as all the tools the hand of the peopple can use this well or not.

 The rise of AI writing tools presents both opportunities and challenges for academic institutions. While these tools can potentially boost productivity and assist researchers in ideation and structure, there are valid concerns around misuse and plagiarism if AI-generated content is passed off as original work without attribution. To uphold academic integrity, universities should implement clear policies defining acceptable AI tool usage, provide thorough training on ethical practices, maintain human oversight and critical evaluation of AI-assisted writing, explore advanced detection methods for improper AI content use, and foster ongoing dialogue among stakeholders to develop best practices. 

There must be the realization that AI is met to compliment the human effort and not to replace the human brain. Responsible usage entails seeing AI an an enabler and not replacer of critical thinking or innovative ability

The algorithm used to conceptualize text via AI and other writing tools is the inputter's. Responsible use of AI tools needs to be implemented at all educational levels. With a transparent description of the algorithms and tools used, we should accept the use of AI for academic writing. It is like the image processing tools we use for wet lab images. Full transparency is needed. 

Like any new tool, AI will be useful for some tasks but not all. When it comes to writing, AI will make the preparation of simple written documents such as memos, scientific abstracts, or intended for internal use documents simpler and less time consuming. For documents such as research papers, AI can assist in preparing some sections of these documents such as introduction, methods, and references, but the more thought provoking sections such as conclusions and discussion will require the personal thoughts of an actual human author. Regardless what it is used for, it needs to be indicated whether AI was used to help generate the document. This requirement should be in place until the time that society evolves to be customed to AI becoming an integral part in the creative writing process.

I am of the opinion that AI tools are being appeared as foe for academic writing. Students of both under and postgraduate levels are being deprived of critical thinking, self-writing and creativity. 

The use of AI in academic writing can be good as well as bad. It can be used for certain tasks well-monitored by human. Otherwise, the results generated by AI can be badly misleading. 

AI-powered resources for academic writing provide many advantages. From ideation to completion, these technologies can help researchers and students at different phases of the writing process. Additionally, these tools frequently offer insightful criticism on grammar and style, which allows for enhanced writing abilities.
In conclusion, there is a complex and wide-ranging discussion regarding AI in academic writing. AI offers previously unavailable chances to improve writing efficiency and increase access to educational materials, but it raises moral and plagiarism concerns that must be carefully considered. Academics may leverage the revolutionary potential of artificial intelligence (AI) while maintaining the integrity and ethics of academic research by embracing a balanced viewpoint.

I truly believe that AI might help in the writing process, but it has to be overlooked by human scientists. I don't think we are at a point, yet, where AI can discriminate between fake news or proper science. Maybe one day, but not yet. Of course, AI can help in the source finding and in the overall organization of the draft process. 

I believe that AI for academic writing is a friend and Not a foe. Research primarily means to come up with algorithms or innovative solutions and not on how that is expressed/communicated. Use of generative AI only helps in expressing the contents in impressive manner. So, as such AI is helping in saving author's time to write the contents and enabling the author to focus more on newer research challenges. 

Given the increasing availability of AI writing tools, how can academic institutions ensure the responsible use of these technologies while maintaining the integrity of research and scholarship? - At student level, use of generative AI tools should not be allowed and the same can be achieved by using AI based anti plagiarism softwares by faculty/admin.

The use of AI for data evaluation and information analysis is not bad and to some extent good due to time-saving and accuracy perspectives. However, the use of AI for writing is considered illegal by most of the institutions/journals. AI is a modern technology that can help to save time with effective information analysis; therefore, I think its use as research labor is not bad. It can help to save time and money with high accuracy and free from human error. However, excess of everything is bad.

While we are witnessing a depressing decline in the linguistic skills of students at all levels, use of AI for writing can only make this decline faster. A person drafting some content with AI, and himself/herself not able to write/speak  2-4 sentences correctly in any language in no way can be expected to be an asset to any society or organization. 

Our efforts must be focused on instilling respect for linguistic skills among student populations rather than being depenedent on AI for writing. If we become too much dependent on AI for linguistic purposes, it will only make us intellectually handicapped.

Language generating AI is at present not performing well enough for high quality academic writing.  I caution using it, unless you know exactly what you are doing.   For example, if you are a junior researcher, you should be careful in using AI for a job application,  as a practiced user can notice when a text is AI generated, which does not give a good impression at all.

Having said that, it can be helpful to reformula a sentence or a paragraph with AI, as long as you know what you want, and you can correct errors,  redundancies, and vague statements in AI generated language.  

Nowadays, artificial intelligence reminds me of the ancient times when the calculator was invented and everyone was doubtful of how it would disrupt students' minds and was prohibited everywhere in schools... Regrettably, it is not within our jurisdiction as educators and professors to decide as it is becoming increasingly challenging each day to identify the excessive use of AI in every single aspect of student work particularly with postgraduate students whom we perceive as abusing the system through AI. 

I acknowledge and admit that AI has incredible tools in making our lives easier and more efficient in academic writing than ever before. But it also presents a double-edged sword and I am not entirely certain how it will enhance our students' progress in research and academic writing. Even with a simple email, we are now hesitant, is this the sender's own writing level or is it that of AI. 

The reality is that no matter what we do using our traditional methods of assessment and evaluation, we will always lag behind. I believe the only approach to address the issue of assessing student performance and the integrity of academic writing is by employing AI itself for this purpose, and most crucial is acknowledging that AI is a reality and is the inevitable future. 

Besides grammar and ortho-typgraphic edition, AI should not have space in academic writting. It use to mask the lack of style or creativity of a scientist to communicate their results just keeps the language stale. If it is used to make a whole section, then it is the AI, not the author, connecting the knowledge in a way that supports (or even creates) findings. Even if it is meant to gather or process large ammounts of data, human revision, a process that tends to take most of the time of the data processing, is required. Otherwise, science would work as a black box of inputs and outputs... It may look like it for people not used to academia (and that is our fault as scientist, and the publishing industry may also be to blame at some extent), but if it is a black box even for us, then science (the understanding of truth) is not science anymore, as we are not understanding anything, just inputing.

Any technology has its way of functioning. Need to agree that on one side, AI writing tools help researchers and students generate ideas, especially when they have no idea.  It helps their writing efficiency. These tools can provide valuable assistance in generating initial drafts based on inputs. But on the other side, to some extent, it diminishes the thinking ability of the researchers and students which finally leads them to issues of plagiarism. Ultimately it depends on the individual researcher/student who uses this tool. 

On one hand, AI writing tools can help researchers and students generate ideas, especially when faced with a blank page. They also can improve writing efficiency. These tools can provide valuable assistance generating initial drafts based on input prompts.

However, there are also concerns regarding the responsible use of AI writing tools in academic settings. These concerns include issues such as plagiarism, the erosion of critical thinking skills, and the potential for bias in algorithmically generated content. To ensure the responsible use of AI writing technologies while maintaining the integrity of research and scholarship.

A balanced approach that embraces the potential of AI while establishing ethical guardrails through policies, education, and monitoring can help maximize benefits and minimize misuse. Responsible AI integration will require an ongoing dialogue between technology developers, academic institutions, and scholars.

AI writing tools in academia offer significant advantages, such as aiding in research, streamlining the writing process, and providing quick access to vast amounts of information. They can enhance productivity and creativity, allowing researchers and students to focus on critical thinking and analysis. However, it's essential to use these tools responsibly, ensuring academic integrity and originality are maintained. Proper guidance and ethical considerations should be integral to their use in academic settings.

The major concerns in AI in scientific publishing are,

Originality
Integrity
Equity
Accuracy
Accountability 
Copyright issues

Read more...
https://storage.googleapis.com/jnl-sljo-j-jccpsl-files/journals/1/articles/8634/64df24fe7f14e.pdf

Ensuring the responsible use of AI writing tools in academic institutions while maintaining the integrity of research and scholarship requires a combination of policies, education, and technological solutions.