Teacher vs copilot
Leveraging GenAI to enhance teaching and learning at a teacher-training institution
PICTURE this: A teacher grades 100 essays in one hour, provides detailed feedback to each student, and identifies class-wide learning gaps — all while sitting under a mango tree enjoying the cool summer breeze. Sounds like a science fiction? Not anymore. When OpenAI launched ChatGPT in November 2022, it didn’t just introduce another tech gadget — it fundamentally challenged how we think about teaching and learning.
As an educator with over 25 years in Jamaican classrooms, I attended my first AI webinar in January 2023, sceptical (as a true scientist) but curious. What I discovered wasn’t just another tool — it was a potential teaching revolution. The question that now keeps me awake at night isn’t whether AI will change education, but whether we’ll use it as a mere digital assistant or embrace it as our copilot, transforming our roles in ways we never imagined possible.
Initially, I treated ChatGPT like Google — a search engine for quick answers. Then I discovered it could create teaching materials in seconds. My perspective shifted dramatically as I explored these capabilities. Today, I primarily use ChatGPT, Claude.ai, Prezi,
Gamma and InVideo.ai (which generates videos from text prompts).
The large language models (LLMs) can, within seconds, create a wide range of educational resources: presentations, interactive multiple-choice questions, game cards, case studies, worksheets, and materials for tests and final examinations. One of the biggest advantages for teachers, I believe, is the ability of the LLMs to assist with grading both formative and summative assessments. Lesson planning, often an onerous task for pre-service teachers, is now more manageable with LLM support.
Use of AI by students
In Shaquille Dunkley’s recent study of tertiary students, most participants used ChatGPT primarily for assignments. Other popular tools include Google’s Gemini, Microsoft Copilot and Claude. I encourage all my students to use these chatbots for lesson plans, study guides, and research support. Importantly, students recognise these tools’ limitations and are trained to verify information using credible sources. With over 25 years of experience, I can spot false scientific claims students might unknowingly include.
Students are required to create presentations as in-coursework assignments and there are many AI tools available to them. The presentations are well structured and appealing. These types of software generate presentations within seconds, leaving time for students to concentrate on content and delivery. In other words, students can hone their public speaking skills with less time spent on formatting presentations.
Data analysis has long intimidated students — some even paid for statistical services. Now GenAI fills this gap. Students no longer need textbooks to determine appropriate statistical tests. Claude.ai, ChatGPT and other tools perform analyses instantly, interpret results, and explain findings. These AI tools demystify data management.
Differentiated instruction — a key component of lesson planning — becomes manageable with AI assistance. Traditional differentiated lesson plans required several hours to conceptualise and write. AI reduces this dramatically, freeing student-teachers to design instructional resources and practise delivery techniques.
Use of AI by facilitators
As a science educator, part of my responsibility is developing students’ pedagogical skills while modelling best practices in teaching at the primary and secondary levels. Tools like ChatGPT and Claude.ai allow me to design resources that effectively support student-teacher training. GenAI allows me to create resources for individual student learners – I can quickly create tiered assessments for diverse learners.
My students are fully aware that I use AI to create teaching resources, mark assessments, and write lesson plans. In the absence of an institutional policy governing AI-based marking, I remain ethically cautious, always seeking students’ permission before using AI to assess their work.
One of the most time-consuming tasks in education is writing lesson plans. I use these tools to teach the structure of lesson planning and encourage students to explore them when creating their own. Students are trained to generate text prompts that will give them the best results. A well-crafted prompt is the “secret to success” of an excellent lesson plan.
Before using a chatbot to mark assignments, it must be trained to interpret rubrics or mark schemes accurately. The more detailed the rubric, the more human-like and consistent the outcome. The advantages of using these models for grading assessments are the speed, comprehensiveness, and depth of feedback — something difficult to achieve under normal teaching conditions. The tools allow for objectivity when giving feedback and students should be able to correct their mistakes before another assignment.
However, these tools still struggle with nuance. In one instance, a student described a solution as “blue-green” after an acid-base reaction. The textbook lists the expected colour change as colourless to “blue”. The AI deducted a mark because it followed the mark scheme strictly, but I chose to override it and award the point. These are the human judgements that machines cannot yet make.
Challenges
The main challenge is cost: ChatGPT Plus and Claude.ai cost US$20 monthly, while InVideo.ai starts at US$35. For educators, the question becomes: what am I willing to sacrifice to reduce preparation time?
Additionally, these tools demand fast, reliable Internet — still problematic in rural Jamaica. Connection interruptions during grading can cause data loss and frustrating file reloads.
There has been much debate about the loss of critical-thinking skills due to over-reliance on these AI tools. However, one of the most exciting features of these models is their ability to generate content and assessments across different cognitive levels of Bloom’s Taxonomy — from simple recall to evaluation. This allows educators to generate resources that will encourage and develop students’ critical-thinking skills.
The way forward
Today we stand at a crossroad. We all feared that calculators would eliminate mental maths, but we embraced it. We must now decide how to integrate AI. It cannot replace us, but it can become our copilot.
We must teacher our students to fact-check AI-generated information and verify sources. Today’s student-teachers need training in AI-enabled classrooms since they’ll shape Jamaica’s educational future.
We must also rethink assessment strategies. Should we reduce marks for lesson plan writing (since AI assists) and emphasise delivery of instruction instead? Should we shift from take-home essays to in-class testing that truly measures critical thinking?
The choice is ours: Will we be the teachers spending entire weekends grading 30 assignments each, scribbling brief comments in margins, falling behind on learning patterns — while family dinners grow cold? Or will we be the educators who leverage AI to reclaim time for what truly matters: inspiring minds, building relationships, and transforming learning?
Dr Audette Bailey is head of the Department of Science at Bethlehem Moravian College.