I have read much in the past week or so about the concern of students and teachers using AI tools, like ChatGPT to work more efficiently. I, of course, advocate for using the available technology responsibly and let’s be clear, I am not advocating students’ nefarious use of technology, i.e. transcribing assignments into ChatGPT and delivering the output as if it were their greatest masterpiece. I do believe, however, that there is the possibility of using AI to get students to think in more sophisticated and advanced ways than they do now. Allow me to explain.
Thinking back to my college days and Bloom’s Taxonomy (still etched into my brain), critical/higher order thinking was
always the ‘holy grail’ (cue the angelic choir). I was under the mistaken impression that moving students from the base to the top of Bloom’s pyramid was a no-brainer - clearly I was very young and naive. Sometimes my students get there, but more often not. This may be a terrible admission for an educator, but I sense it is more common than folks in education would like to admit.
As I delve deeper into the realm of AI, continuously learning and creating, I contemplate whether crafting prompts for AI can enhance students' capacity for critical thinking and help them overcome the figurative barrier that, in my perspective, lies at the mid-point of Bloom's Taxonomy, impeding their progression to higher-order thinking.
I have always thought you haven't truly learned until you can take an idea apart and put it back together as something different - making an idea or concept your own. At least that’s what I tell my students. But it is rare for students at the middle and high school level to develop something that is truly their own original idea or product. Don’t get me wrong, some students can, but this is allusive to the masses in my experience.
As educators, we are often reminded that engagement plays a crucial role in learning. When we connect our teachings to real-world applications or discover topics that ignite students' passions, they become more likely to stay engaged in the learning process over an extended period. This, in turn, increases their chances of achieving genuine and lasting learning outcomes.
The 'Holy Grail' of Education
The ability to Analyze, Synthesize, Evaluate, and Create - this is critical thinking.
What is more real-world and exciting than technology and AI? Nothing. So how does AI fit with getting students to the next level in their thinking?
They must deeply understand a concept or idea:
How to analyze it - take it apart, consider it from different perspectives, and how do the parts work together
How to synthesize it - connect it with other things they know - put it back together, understand how the parts fit together.
Feel confident enough to make judgements and form well-reasoned opinions about a concept or topic.
And finally, creating something original be it a thought, concept or product.
According to Elder and Paul (2010), critical thinking includes:
Raising questions and problems, communicating them clearly and precisely
Gathering and assessing relevant information
Come to well-reasoned conclusions and solutions
Thinking open-mindedly within alternative systems of thought (assumptions, implications, consequences)
Communicate effectively with others in figuring out solutions to complex problems
This is not rocket science and I expect there are few who disagree with the competencies put forth by Elder and Paul. However, I am going to go out on a limb here and posit that one way students can get to the upper echelon of thinking is through AI and effective prompt engineering. Prompt engineering involves structuring requests (prompts) in a way that can be understood by AI to produce a desired outcome. The complexity of the output determines how in-depth the prompt must be.
Where the magic can happen
My engagement with AI has been a self-initiated learning journey in its own right. It demands time and persistent effort to contemplate, create, and often iterate multiple times to achieve the desired outcome. One must engage in profound thinking to formulate the right questions for the intended results. From my experience, it's not a one-time task but rather an evolving process that involves continuous refinement and repetition until the desired output is achieved.
Let me give you an example:
One of my initial prompts stemmed from a common issue students faced when tasked with writing about a topic of their choice - they often felt 'stuck' even before they began. To help them brainstorm, I typically posed similar questions to prompt their thinking. I started pondering whether AI could provide valuable assistance, potentially saving time. This concept eventually evolved into the Writer’s Idea Generator. The following outlines my iterative process:
1st attempt: (no clue what I'm doing) “tell me what to write about” - this generated a laundry list of potential topics (not exactly what I was looking for) - too general.
2nd attempt: “help me write an essay” - this generated a list of questions about topic, type of essay, length, etc.
This level of thinking, it’s very basic, nothing specific and in terms of Bloom’s taxonomy this is definitely the base of Bloom's pyramid, but I am learning as I go. As my attempts continue, I begin to realize that I need to drill down into what specifically I want the output to be. At this point I also find that what I really need is to create a tool for students and not just have a conversation with ChatGPT. After some research, I decided to develop a custom GPT that others can use (after all this is really what I need). Seeing the error of my ways, I continue:
3rd attempt: “help the user decide on a topic for a writing task. Begin by asking what type of writing they will be working on narrative, persuasive, or argumentative.”
I could see this was the start of something good, but how will this get the user what they need?
4th attempt: “help the user decide on a topic for a writing task. Begin by asking what type of writing they will be working on narrative, persuasive, or argumentative. Then ask the user about their interests.”
This initial prompt generated a set of questions categorized into seven themes, covering personal experiences, travel, relationships, and more. However, I soon recognized that my ultimate goal was to develop a tool that students and others could readily utilize (which was the initial motivation for my journey into this domain). Consequently, I shifted my focus from merely engaging in conversations with ChatGPT to creating a customized GPT model. This transition necessitated research, multiple iterations, and, eventually:
The Finale
“You are a writing coach with specific knowledge of narrative, argumentative and persuasive writing. Students come to you to help them generate ideas for what to write about when they have writer's block.
Start by asking the user what type of writing they would like to produce: narrative, argumentative, or persuasive writing. Ask the user about their interests, curiosities, and any other questions to help the user find an acceptable topic for their writing. Make writing topic suggestions to the user based on their responses.”
I welcome you to try this custom GPT model for yourself on the Try AI page. This model is quite basic, but involved analytical, evaluative, and ultimately the creation of an original work! Let’s be realistic, not everything can boil down to a prompt, but I believe there is merit in looking at prompt engineering as a potential vehicle to get students to higher order thinking.
This is how we shift the paradigm. I welcome your thoughts, comments, and critiques.
Enjoying my Blog? Please share this site with colleagues and friends and subscribe to be notified when new posts are available. I encourage you to reply with comments or content suggestions.
Elder, L., & Paul, R. (2010). Critical Thinking: Competency Standards Essential for the Cultivation of Intellectual Skills. The Journal of Developmental Education, 34(2), 38–39.