As an Associate Professor in Health Sciences and Faculty Academic Assessment Director
at California State University Channel Islands, I have recently integrated Chat-GPT, GPT-4,
DALL-E tools into my work, witnessing firsthand their transformative potential to enhance
productivity in my busy life juggling multiple responsibilities. I conduct research, teach,
hold an administrative role as Academic Assessment Director for the University, raise two
elementary-age boys, and am a first responder’s wife. While I am a social worker by training, I
am a millennial who grew up adopting new technologies. Growing up, I was the person to “fix.”
the home computer if anything went wrong. Whether or not you’re an early adopter of new
technologies, I hope you can see how the ideas below may improve your productivity in
your academic life. Here, I summarize my early experiences and the multifaceted applications of
these AI models in academic settings.
Streamlining Cross-Sectional & Longitudinal Pre-Post Data Analysis
In my past year as Academic Assessment Director, I promoted the use of Canvas
Outcomes as an assessment tool that meets faculty where they are in Canvas. The tool creates
spreadsheets of raw data on student assessments for programs and the university to track
assessments of students over time. One of the most time-consuming aspects of assessment and
research involves the analysis of longitudinal datasets. Before trying GPT-4, I had to clean and
combine datasets to compare pre-post data. With the use of GPT-4, I could upload pre- and
post-datasets separately and describe the analyses I desired, specifying which variables in the
dataset I wanted it to compare and how to match students across the datasets. GPT-4 provided the
results of comparing mean scores for students in a pre-post test within seconds. I prompted
it to create tables and a summary of the results, which it also did within seconds. Of course,
we must interpret the results ourselves. GPT-4, with the ability to upload data into the application
and its advanced computational skills, has significantly reduced the labor involved in
analyzing pre-post data from student assessments. I also apply this to my research in helping
hospital and non-profit community health organizations analyze pre-post intervention data.
Enhancing Idea Generation and Program Planning
The task of generating new ideas and developing academic programs can be daunting.
However, using GPT-4 for idea generation from faculty notes has proven to be a game-changer.
In one instance, GPT-4 analyzed two years’ worth of competency ideas from a Google
spreadsheet for our Health Science program, employing a Cultural Taxonomy framework to
create Program Learning Outcomes (PLOs). We uploaded our faculty notes spreadsheet
document and Cultural Taxonomy document; we prompted GPT-4 to create six ideas for
Program Learning Outcomes based on the ideas document. See my prompt and GPT-4 response. The PLO ideas it gave us within seconds were an excellent starting place for our team
discussion. We loved four of the PLOs that GPT-4 developed but didn’t love the other two. We
revised the four we liked and created two new ones from scratch. The AI’s ability to sift through
extensive notes on competencies and learning outcomes to create draft PLOs within seconds has
revolutionized the way we can approach program development and assessment.
Facilitating Sample Size Analysis
Determining the appropriate sample size for assessment projects and research studies is
critical for ensuring statistical power and validity. GPT-4 has been instrumental in conducting
sample size analyses, offering precise recommendations that align with the research objectives
and constraints. After prompting GPT-4 to conduct the power analysis, I asked it to “show me its
work” and break down the statistical formula that it used. I confirmed that it was using accurate
numbers in the analysis. I conducted the analysis by hand and obtained the same result,
confirming the accuracy of GPT-4 to perform this function. This capability has greatly assisted in
the planning stages of studies, ensuring that they are designed with rigorous scientific standards
in mind.
Data Visualization and Interpretation
GPT-4 can read and create pictures and charts. Oftentimes, we want to have a visual
image representing our data to observe trends. I have prompted GPT-4 to create different types of
visual representations of data, such as scatterplots. I was able to upload the data, prompt the AI
tool on what type of chart to create, and specify the colors used in the chart. By providing
context for the data, these AI tools can generate insightful visualizations that elucidate complex
patterns and trends, facilitating a deeper understanding of the underlying narratives.
Timeline Creation
We create timelines for students’ progress in our courses and/or program curriculum.
Almost every time we apply for funding for our research, we are asked to provide a timeline.
Often, we discuss our timeline in the syllabi for courses and narratives that we create describing
our studies. I uploaded a narrative description of a study I was proposing for a grant application
and asked GPT-4 to create a timeline. I informed GPT-4 where I was submitting the grant. It
created a detailed timeline from the information I had already referred to in my narrative grant
application. I revised the timeline, but it provided me with a starting place. At first, the timeline
was written in text. After revising it, I uploaded the text timeline description into GPT-4 and
asked it to create a visual representation of the timeline. I added it to my grant application. GPT-
4’s ability to create detailed timelines has helped in planning and executing projects more
effectively, ensuring that milestones are met and progress is systematically tracked.
Image Creation
OpenAI, the company that created GPT, also created DALL-E, an AI illustration, art, and
visual creation application. I used it to create the accompanying picture used in this blog. I have
used it to create images for class and for my research. You can prompt it to create any type of
image that you would like. If you are prompting it to create images of humans, I have learned
that you can be specific down to the hairstyle, hair color, clothing, and background color or
design. I have used DALL-E to create images to use in my research presentations as well.
Voice, Reading Level, Grammatical & Typographical Changes
You can use GPT to change the tone, voice, reading level, or style of writing, as well as to
fix grammar and other errors. I uploaded this blog into GPT-4 and asked it to fix any
grammatical and typographical errors. It sent a revised version of this blog back to me. I have
asked GPT to fix the tone of my writing from passive to active, as I have been informed that I
tend to be a passive writer in the past. I strongly believe these methods of using AI to improve
our writing can help people who want to improve their writing or shape their writing for a
specific purpose.
Ethical Considerations and Data Privacy
While the benefits of integrating Chat-GPT and GPT-4 into academic research and
service are immense, it is crucial to navigate these advancements with a keen awareness of
ethical considerations, particularly regarding data privacy. For example, if faculty are analyzing
student or research participant data, they should de-identify the data before using GPT for
analysis. The option to enhance data protection by opting for a paid model of GPT-4 (called
Team or Enterprise subscriptions) that ensures the privacy of entered data is a step in the right
direction. We need ongoing vigilance and ethical stewardship in the use of AI tools in academia.
Conclusion
The integration of Chat-GPT and GPT-4 into academic research and service has opened
new horizons for efficiency, innovation, and depth in scholarly work. From data analysis to
program development, these AI tools have proven to be indispensable assets. As we continue to
explore their potential, it is imperative to balance the benefits with a commitment to ethical
practices and data privacy. The journey of integrating AI into academia is ongoing, and it
promises to reshape the landscape of research and academic service in profound ways. I am
integrating AI into my teaching weekly this semester and look forward to sharing my lessons
learned in the future on AI and teaching.
I would be happy to show you how to use GPT and DALL-E. Email me, Kristen Linton, MSW,
Ph.D. Have fun exploring these AI tools.
