Why all the buzz about AI-assisted writing? Think about it—haven’t we already embraced tools like Grammarly and Quillbot and other AI-assisted and Computer Assisted Writing to help us write better(Wang, 2022)? And remember when we switched from digging through library cards to hopping onto research databases? Evidently, each has advantages and disadvantages (Falagas, 2008). Sure, there was a time when many educators were wary about students using computers for writing, worried it might spoil their writing skills (Billings, 1986) or second language acquisition (Lai, 2006; Gündüz, 2005). But look how that turned out: we adapted and learned to see the value in the technology. So, what's the big deal now? AI writing tools are just the next step. Instead of pushing back, why not dive in, learn how it works, and show others how to use it? Let's make the most of what tech can offer and keep up with the times!
Billings, D. M. (1986). Advantages and disadvantages of computer-assisted instruction. Dimensions of Critical Care Nursing, 5(6), 356-362.
Falagas, M. E., Pitsouni, E. I., Malietzis, G. A., & Pappas, G. (2008). Comparison of PubMed, Scopus, web of science, and Google scholar: strengths and weaknesses. The FASEB journal, 22(2), 338-342.
Gündüz, N. (2005). Computer-assisted language learning. Journal of language and linguistic studies, 1(2), 193-214.
Lai, C. C., & Kritsonis, W. A. (2006). The advantages and disadvantages of computer technology in second language acquisition. Online Submission, 3(1).
Wang, Z. (2022). Computer-assisted EFL writing and evaluations based on artificial intelligence: a case from a college reading and writing course. Library Hi Tech, 40(1), 80-97.
We've come a long way, from libraries to research databases, but researchers still have to put in the work. It requires considerable mental effort to find relevant articles, sift through them, and extract the necessary information.
But in contrast, students can simply put their homework questions into an LLM and get well-crafted responses immediately. They can then claim this generated text as their own brainchild without even thinking about the topic in question a single time.
By not caring if young students are simply regurgitating what GPT answered to their homework or exam questions, you are basically robbing them of their precious developmental years. Critical thinking and problem-solving are essential skills for every human, and we develop those skills by researching, evaluating information, and synthesizing it.
When I wrote essays as a student in high school, I remember that sometimes I would run out of ideas of things to say about a topic. To remedy the writer's block, I would close my eyes and think about everything I knew about the subject, creating connections between that knowledge and any type of other prior knowledge that may be relevant to try to come up with an interesting point to add to my essay.
This process was mentally taxing, almost like trying to remember a dream as you wake up. If I had access to GPT and wanted to use it for this, it would give me a thorough and well-structured essay in seconds. However, as you might have guessed, the assisted answer would in no way help me develop my cognitive skills, imagination, and creativity. Memory levels would also take a huge blow because memory is also a skill that requires practice.
I agree when you say that we should embrace AI (AI has been around since the 50's by the way), but we need to know how to use it efficiently and safely. It's simply a (very powerful) tool, and we should be using it as such.
The most efficient way to use LLMs, in my opinion, is to regard them as an auxiliary resource. They can assist in brainstorming, clarifying ideas, or fact-checking, but the main intellectual labor still needs to come from the individual.
Assisted writing by AI has been getting a lot of attention in recent years for several reasons:
However, there are also concerns about the overreliance on these tools, especially regarding their dependence for tasks that could benefit more from human creativity and sensitivity. Additionally, ethical issues such as authorship attribution and data privacy are relevant when it comes to AI-assisted writing.
One way that AI-assisted writing and become better is that we use it for semantic searches. I find this problem so frustrating that everything I want to know get an answer while researching, I have to go through the same pain of Googling and then reading articles, etc.
This is because although their's tools like ChatGPT, there are 2 problems -
1. their knowledge is cutoff from certain period
2. they hallucinate and have no reliability that where there answers comes from..
I guess there's a need for an AI-based alternative for Google Scholar.. and that's what researchers need at this moment.
The main concern surrounding AI-assisted writing is transparency. Authors should have the freedom to disclose that their work is AI-assisted without facing stigma. Non-disclosure, on the other hand, can be problematic for the reasons already mentioned by other commenters.
Another important aspect is the distinction between assistive and generative AI. The ethicality of assistive AI, which aids in tasks like grammar correction and style suggestions, is relatively accepted. However, the use of generative AI, which can produce entire pieces of text, remains a topic of hot debate.
I recommend that authors stick to assistive AI, and not generative AI.
For full disclosure, this post is an example of AI-assisted writing (non-generative).
The same can be said about AI-assisted research. Just as we moved from library cards to online libraries and databases, the use of AI-assisted keywords by abstracting and indexing agencies like Scopus and WoS is both legal and advantageous, isn't it?
Embracing these technologies not only helps us adapt but also allows us to maximize the benefits they offer. Let's continue to explore and integrate these tools to enhance our work.
Dear El Mehdi Seddougui
Thank you for sharing your thoughtful perspective on the role of AI in education. I wholeheartedly agree with your points about the importance of critical thinking and problem-solving skills and the potential risks of over-relying on AI tools for writing and research tasks.
The same technology is at the disposal of teachers, provided they are willing to keep up with it. The same AI that produces the writing can also be used to detect it. Additionally, we can resolve concerns about students simply regurgitating AI-generated content by requiring them to give interactive presentations on their papers. If a student can effectively support and explain the ideas in their paper, it means the teacher's goals have been accomplished.
Ultimately, how students reach their goals should not be our primary concern. With the current rate of technological advancements, our role as educators will increasingly shift towards guiding and directing students rather than traditional teaching. By embracing AI as an auxiliary resource, we can better prepare students for a future where these technologies play an integral role.
Thank you again for your insights. I look forward to continuing this important conversation.
Best regards,
Dear Guilherme Semionato Galicio ,
Thank you for bringing up the important concerns regarding overreliance on AI tools, particularly in tasks that benefit from human creativity and sensitivity. I completely agree that ethical issues such as authorship attribution and data privacy are crucial when discussing AI-assisted writing.
Could you please expand on the ethical considerations? Every time I have spoken on this controversial matter, the first question has been about ethical considerations. However, I don't see anything inherently unethical about getting help from AI search engines or AI-assisted writing, as long as the author takes full responsibility for the entire piece. Could you elaborate on the exact ethical issues you have in mind?
Your insights on this would be very helpful for a more nuanced understanding of the ethical dimensions of using AI in writing.
Best regards,
Dear Udit Raj,
Thank you for your insightful comments on the current limitations and potential of AI-assisted writing. I completely agree that using AI for semantic searches could significantly improve the research process.
We are just at the beginning with AI, and these complaints and disagreements with AI actually help its growth and accuracy. With paid versions and additional AI tools accompanying ChatGPT, access to the most recent articles is no longer a problem. However, as you mentioned, the replies can sometimes be imaginary or unreliable. This is where human intelligence becomes crucial. Double-checking with Google Scholar and reading the suggested papers ensures the accuracy and reliability of the information. Always ask for citation references or links to the article. You will find out that, in some circumstances, the link and the article do not exist.
Remember, you can engage with the results and demand more accuracy from AI tools. As time goes by, you can expect more accurate answers and improvements in AI capabilities.
Best regards,
Dear Imam Salehudin ,
Thank you for your thoughtful insights on the transparency and ethical use of AI-assisted writing. I agree that authors should have the freedom to disclose their use of AI without facing stigma, as non-disclosure can indeed lead to various ethical concerns.
The distinction you make between assistive and generative AI is particularly important. The ethicality of assistive AI, which helps with tasks like grammar correction and style suggestions, is generally well-accepted. In contrast, the use of generative AI, which can produce entire pieces of text, continues to spark debate.
At what point do we move from assistive to generative?
The choice between assistive and generative AI is not always clear.
I hope my colleagues will further elaborate on the "why" reasons for transparency in writing and the "when." Don’t we pay for editorial services recommended by journals to help with the writing? Is that assistive or generative? We do not give any credit to these editorial services in our papers, yet journals often highly recommend using their services. Is this ethical? Some even ask for proof of editorial services before moving into publication. Is that ethical?
Thank you for clarifying that your post is an example of AI-assisted writing (non-generative). This transparency is indeed crucial for fostering trust and understanding in discussions about AI in writing.
Best regards,