Is AI Your Finish Line or First Step?
If you asked 10 teachers if AI saved them time, you’d get 12 different answers.
Truthfully, how effective AI is for teachers isn’t up to the AI as much as it is where in the workflow the AI gets used.
Let me explain.
Here is a simple workflow that will lead to frustration and failure:
1) Have an idea for a resource.
2) Prompt an AI to create the resource.
3) Finished.
This workflow assumes that AI should get me straight to the finish line in my work. It assumes that AI should give me a lesson plan, rubric, or quiz that will be perfectly suited to my learners, my context, and my teaching style.
These are poor assumptions.
When we expect AI to do our whole job for us and produce classroom-ready resources, we’re often disappointed.
When I was consulting at a school recently, a student told me that the week before, they’d sat a 20-question multiple choice quiz. Every correct answer was ‘all of the above’. Her teacher clearly thought AI would get him to the finish line.
What should that teacher have done instead?
Treat AI as the first step, not the finish line.
Treating AI as the first step acknowledges that we are the professionals. While AI might get
us going in the right direction, no resource will be ready for our classroom until we’ve given it the all-clear.
The ‘AI-as-first-step’ workflow looks like this:
1) Have an idea for a resource.
2) Prompt an AI to create the resource.
3) Iterate on AI response with AI, considering your professional context.
4) Perform final tweaks by hand.
5) Finished.
As teachers, we have
- Intuition
- Discretion
- Familiarity
- Experience
- Pastoral instincts
- Content knowledge
that an AI simply doesn’t have.
Using AI as a first step allows us to leverage its benefits without cheapening our own practice as educators.
There are two parts of using AI as the first step:
1) Refining with AI
When you get you’re initial output, it will most likely not be fit for purpose.
Going back and forth with the AI to make the work more suitable is called iteration. This is a crucial step. You’ll often give the AI instructions such as:
Please rewrite in fewer words.
Please clarify the meaning of _______.
Please rewrite this in less formal language.
You missed one part of my prompt; don’t forget to _____.
This is a crucial step in ensuring the resources you create are fit for your classroom, context, and learners.
The final step involves moving on from AI use.
2) Performing final tweaks by hand
For argument’s sake, say you were creating a multiple-choice test with AI.
17 of the 20 questions may be great.
In my practice, you could keep trying to iterate until you’ve got 20 perfect questions, but you might be there all night.
Sometimes, it’s best to move the process out of the Large Language Model and perform the final tweaks by hand.
The sort of things that may be easier to do by hand are:
Making a specific adjustment for one of your learners.
Linking to a piece of content you addressed in a previous lesson.
Integrating your school’s vision/mission/values/learner attribute.
Integrating a specific religious/ethical idea.
The best work is done with a combination of Artificial Intelligence and Organic Intelligence.
Conclusion
How effective and efficient our AI use is comes down to where we have it in our workflow.
Remember, AI is the first step, not the finish line.
This Week’s Recommended Resources:
ARTICLE: AI could be the key to targeting students’ individual learning needs
https://news.csu.edu.au/latest-news/ai-could-be-key-to-targeting-students-individual-learning-needs
YOUTUBE VIDEO: AI 101 for teachers: demystifying AI for educators
https://www.youtube.com/watch?v=DosAhBSjNds
ARTICLE: With generative AI we can reimagine education — and the sky is the limit
Until next week,
Happy Teaching!
Paul Matthews
