|Photo by Ricardo Gomez Ange|
Contributor: Dr. Marian Kennedy
Writing is simply assembling coherent words into text and we humans have been doing it for a while, since at least 3400 B.C. While writing teaching begins (or anyway should) in elementary school, many people hone it for their entire lifetimes. Indeed, I still stumble across rules for technical writing as I develop or that are new, even though writing is a major component of my professional responsibilities as a university faculty.
As graduate students in a STEM field, you are training for a profession where you will be asked to develop proposals, memos, manuscripts, news releases, etc. To make sure you are ready, faculty in your department have designed curricula that will inevitably require you to develop well written documents such as literature reviews, candidacy proposals or a manuscript(s) suitable for publication in peer-reviewed journals. In addition, your campus may also provide you access to writing centers that provide feedback and publish lists of best practices for scholarly writing on their graduate school websites. Like Clemson, they may also provide you access to an artificial intelligence (AI) writing assistant. Grammerly Premium checks for grammatical errors and makes other writing-related suggestions. If you have access, this tool decreases the barrier for getting help- you can instantaneously get feedback on demand with no need to schedule an appointment. It does indeed help with the professionalism of my writing, but I am sure that did not change my life. I still need other humans, such as research collaborators and editors, to provide additional feedback on the document grammar and spelling along with comments on the ideas included (or missing) from the manuscript text. Grammarly rises the quality my writing beyond that of Microsoft Word’s grammar check tool, and therefore decreases the slog of an editor through a draft.
Since I had never tried one of these tools to ‘write’ professionally, I decided to try Moonbeam. Pro tier is currently $15 month; we stuck with their free tier and used their “Ask Luna” feature to create text related to the topic, “self-care for graduate students”. In a few seconds, it provided us with a clear paragraph on the benefits of self-care during graduate school (lower stress) and six self-care tips (establish boundaries, take break throughout the day, exercise, connect with friends and family for emotional support, make time for self-care activities like medication and seeking professional help). All of these six tips are valid, and four of them appeared in my post on the subject. Moonbeam was much quicker than we were! So, my takeaway was that Moonbeam was a good tool to help me as I write general pieces. To ensure that I have an appropriate scope for topics where there has already been a vast amount written (such as time management), I will continue to use Moonbeam. In this post, I used it to make sure that I have provided you with a larger number of AI writing assistant options. (You can see the additions suggested by Moonbeam below (in red) below that were missing from my original draft). But I remain skeptical that I could use this when drafting a research paper where writers have not yet generated enough content that could be found or used by the AI tool.
But what are the ethics of using AI for your writing projects? Sunaina Singh tackled this on her blog, Paperpal), in the post entitled “The ethics of using AI in research and scientific writing She highlighted the need to disclose when you created an article with the assistance of AI, especially when the tool’s creative input is high. Depending on the type of assistant, the disclosure could be highlighting that the piece was cowritten with AI or as an acknowledgement if the AI input was not critical to the article’s intellectual findings (such for a scientific research paper).
If you have not yet tried an AI writing tool, here is a list of options that you may want to try:
- ChatGPT: This system emerged in 2022 from OpenAI and is distinct from GPT-3. GPT-3 was the third generation of OpenAI's GPT language model while is a variant of the GPT-3 model specifically designed for chatbot applications. It has been trained on a large dataset of conversational text and is also generally faster and more efficient than GPT-3.
- ElicitUsed to find relevant papers and summarize takeaways from a journal paper specific to your question.
- GPT-3: Generative Pretrained Transformer 3 is an AI-powered language model developed by OpenAI. It can generate human-like text from a few words or phrases.
- Grammarly: Checks writing for grammar, spelling, and punctuation errors. It also can detect plagiarism.
- Hemingway Editor: Checks for grammar and spelling errors, and can also suggest ways to simplify and improve your writing.
- lex.page (aka Lex): This tool is powered by GPT-3 and helps you generate essays, articles, etc. While I was put on a waiting list to use it, there is a good summary here.
- Moonbeam: With a prompt, this tool can generate content or suggest new ideas. This was the tool tweeted about by Brian Albrecht and I tried while writing this piece.
- ProWritingAid: Checks for grammar and spelling errors, as well as offering feedback on writing style, tone, and clarity.
- Quill: Can help writers generate content quickly and accurately. It can also suggest new ideas and topics to explore.
Closing: Have you used AI to assist you during your writing? What AI tools have you tried? Share your experiences in the comments below.
Acknowledgements: This blog post was edited by Kate Epstein
of EpsteinWords. She specializes in editing and coaching for academics, and
she can be reached at kate at epsteinwords.com. Many thanks to Moonbeam for insightful comments and to my spouse for suggesting this topic. He took screenshots of Dr. Albrecht's texts since I have limited my social media consumption only to LinkedIn.