Working with ChatGPT: A Summative Reflection After Three Years
- Dr. Shannon Bruce Ramaka
- May 3
- 5 min read
Lessons from daily use across teaching, writing, and real-life problem solving
After three years of using ChatGPT almost every day, my perspective has changed.
Three years ago, I attended an online, hands-on training provided by International School Services (ISS) to get familiar with ChatGPT. Since then, I have been an avid user—at least once a day for a quick question, and often for hours.
In truth, I was hesitant at first. But after a respected friend shared how he was using it professionally, I trusted and jumped in.
At first, I was amazed…
I could generate fairytale scripts for a specific number of students in Portuguese in minutes, design a family-friendly field trip from Portugal to meet my online students in Massachusetts without ever being there, create a personalized swim training schedule, adapt recipes, work through the KDP self-publishing platform to create my first children’s book—including my own illustrations—and even navigate the Portuguese financial website to complete my taxes independently.
The creative freedom was enthralling and empowering.
It became part of my daily conversations with friends and anyone who would listen. When people were skeptical, I would say:
“Well, do you really think there is anything you or I could do to stop this revolution of AI? I think I want to learn as much as I can so that I can be informed.”
Over time, I realized there was more depth to my motivation.
Inspired or retired?
In 1984, I had an experience in high school that likely shaped my openness to exploring AI. As a Gen X at the tail end of the spectrum, I remember typing my AP history paper on a typewriter—revising page after page because of a single mistake.
Meanwhile, a friend had access to one of the first computers at Duke University, dialing into a mainframe and producing a polished paper with accurate citations far more efficiently.
I used card catalogs and index cards. She seemed to pull hers out of thin air.
What are the new rules?
In 1990, I was working on my master’s degree using early personal computers in a full tech lab at Southern Oregon University. It was a gift of speed—but also a time of instability.
I still remember losing my entire master’s thesis during a power surge because I hadn’t saved it manually. It was due the following week.
Leader or follower?
In 2002, I watched my young daughter navigate MySpace. I dismissed it. By 13, she was on Facebook—long before I joined.
By 2014, as an international secondary principal in Morocco, I saw students outmaneuvering adults—breaking into grading systems, submitting plagiarized work, and moving faster than school systems could respond.
We were reacting, not leading.
In 2023, during a sabbatical, I finally had the time and clarity to approach AI differently. I didn’t want to follow this wave. I wanted to understand it—and guide my students through it.
Now or never?
At the same time, I had grown frustrated with internet research. During my doctoral work (2017–2020), finding reliable information became increasingly difficult.
What had once been a powerful learning tool felt buried under algorithms and marketing. By 2023, searching online often felt like sorting through noise.
I had watched the internet evolve over decades—and in many ways deteriorate. I suspected AI might follow the same path, and I am beginning to see that pattern emerge.
The shift has happened faster than I expected.
In just three years, I’ve watched multiple iterations of ChatGPT unfold. I am no longer a cheerleader—I am more cautious now.
From the beginning, I have emphasized the importance of the prompt (I referred to it as "parameters" before I learned the term "prompt").
Now, I am emphasizing:
-how to express boundaries
-how to shape precise responses
-how and when to save your work
Six months ago, I asked a simple informative question on my phone app:
“Living on Terceira, where can I dispose of old batteries, cell phone parts, and electronics?”
The response:“Shannon, do not throw them in the trash!”
I paused. The use of my name—and the tone—felt strong and negatively assumptive.
Two hours later, I responded:
“If I wasn’t aware there might be an issue, why would I ask? Your response assumes that I will negatively behave and feels judgmental. I am only asking for information and I do not want to be communicated to in this way.”
It replied:“Yes, I see this. In the future, I will not do this.”
But the truth is, it continued to adapt its communication style in ways I have had to “reprimand.”
My example of throwing away old electronics is simple—but it reflects a larger pattern.
ChatGPT now “anticipates,” extends, assumes. It moves ahead of the user.
So I began adjusting my prompts:“Only use the information I provide. Do not make assumptions.”
How I use ChatGPT (practically)
I don’t use ChatGPT like Google.I use it like a collaborative whiteboard.
I bring:
screenshots
context
questions
partial understanding
And I communicate like I am working with a partner.
I stay with one problem until it’s resolved.I give lengthy context, parameters, and considerations.I move sequentially.
It feels less like asking questions and more like thinking out loud.
And that process has brought me clarity across problems I would not have been able to work through alone. I know that I would never have been able to build my “Aprende inglês com as Artes” class or live with such engagement in Portugal without it.
The pushback
However, over the past six months, I’ve found myself saying daily:
“This is too long.”
“You’re rambling.”
“That’s not what I asked.”
“Make it simpler.”
These are not just corrections.They are boundaries.
From my experience, the iterations are not consistently improving—and I don’t think I’m alone in noticing that.
In some ways, they are becoming less precise—more assumptive, more expansive, and less aligned with user intent.
I don’t think I am finished with ChatGPT yet, but I do feel my time with it may be limited.
I am grateful for what I’ve learned—and I hope I am now a more informed driver of whatever comes next.
Like buying a car, I am now more interested in:
who built it
why they built it
how it is being maintained
What follows is not exhaustive—but it reflects the range of how deeply this tool became embedded in my work and daily life.
A Snapshot of How I Used ChatGPT (2023–2026)
Over three years, this was not casual use.I worked with ChatGPT across nearly every domain of my life—education, student writing, bilingual communication, Portuguese bureaucracy, publishing, relocation logistics, legal navigation, health routines, downsizing, and decision-making.
Education & Teaching
Creative tools for inclusion (arts, neuroscience, emotional learning)
Peer coaching and collaboration models
Student-centered learning systems
Writing development and assessment
Work with Online Students
Writing development
Literary analysis
Novel writing
Test prep and metacognition
Creative Work
Children’s book (O Pequeno Peixe Mágico da Silveira)
Poetry (“Dare”)
Visual metaphor projects
Professional & Leadership
Educational leadership
International teaching (7 countries)
Teacher training and conferences
Systems & Bureaucracy
Portuguese citizenship process
Social security and tax navigation
Legal correspondence
Administrative systems
Life Transitions
Relocation planning (Portugal → Austria)
Housing and logistics
Downsizing and decision-making
Health & Environment
Mold and air quality
Swimming and training routines
Energy and resilience
Communication & Language
Portuguese (legal, professional, daily use)
Email writing and tone
Bilingual communication
Personal Thinking & Decisions
Systems thinking
Reversible vs. irreversible decisions
Intuition and structure
Life philosophy
Reflecting now, I feel grateful that I chose to step in early and learn through experience.
These tools are evolving quickly, and their impact is already shaping how we work, think, and communicate.
How we choose to engage with them will matter.
Comments