What is Prompt Engineering?
As artificial intelligence increasingly makes its way into education, “prompt engineering” emerges as a critical skill for educators seeking to harness AI’s potential to unlock learning for all students. Especially for those with dyslexia and other learning disabilities, carefully crafted AI prompts hold the key to providing the personalised literacy instruction long promised but seldom delivered at scale in traditional classrooms.
Prompt engineering refers to the strategic guidance given to AI systems to produce relevant, helpful responses. Like an ideal teacher differentiating lessons based on individual needs, prompt engineering provides the context and parameters for AI to generate custom materials for unique learning profiles. For the dyslexic learner who struggles with decoding, a precisely tuned prompt can produce tailored, accessible texts and exercises promoting reading mastery.
Teachers have long dreamed of a classroom where individualised education plans (IEPs) could truly individualise. Now with AI, this vision edges closer to reality. Yet without the skill of prompt engineering directing AI capabilities, machine learning on its own cannot fulfill the promise of personalisation. Structured prompt frameworks control for lexical complexity, phonetic components, vocabulary depth, topic relevance, and more—all targeted interventions for unlocking potential in struggling students.
In essence, prompt engineering operationalises the link between learning science and AI capabilities, cascading differentiated instruction to all learners. As schools explore responsible and equitable applications of AI, skilling teachers and students in prompt engineering emerges as a priority for serving those with learning disabilities. Equipped with carefully honed prompts, AI could finally fill the instructional gap holding so many bright young minds back.
What this looks like
If you’ve been reading the AutSide for a while, you’ll know that I like to play around with ChatGPT and other LLMs. I’ve begun experimenting with it to differentiate readings for my Developmental Reading Learning Centre, which features a wide and complex blend of learning gaps and needs. With AI, and prompt engineering, I can now create very specific, very tailored readings and word banks for my students. I do this by describing the student in as much detail as possible, then asking the model to generate a prompt that will in turn be used to create the materials.
So, first, I would collect data on the specific needs of each student. This could include information on:
Their current reading level
Areas where they struggle (e.g. phonetic decoding, sight word recognition, fluency)
Topics they are interested in
Learning preferences (visual, auditory, etc)
Then, I would feed this data into ChatGPT. The algorithm is programmed with rules to control vocabulary, sentence length and complexity, phonetic patterns, word frequencies, and other attributes based on the individual student's needs and interests.
In addition, it can automatically generate custom word lists for each student focusing on their specific areas of difficulty. For a student struggling with "ough" vowel teams, it may produce a list of words containing "ough" to practice. For another student working on silent letter combinations, it would generate words with those patterns.
The AI-generated reading materials and word lists can be directly incorporated into lesson plans and assignments. And because the AI can produce these customised learning materials instantly, it would save teachers substantial time they’d otherwise spend finding or creating such resources manually. Teachers could then dedicate more time to working directly with students.
The end result is an automated way to provide each learner with targeted reading content catered to their needs, while freeing up the teacher to focus on instruction and guidance.
An example
Prompt: William is a 17 year old student with a 1100 lexile level, eigth grade vocabulary, and 7th grade comprehension skills who struggles to accurately decode texts at his grade level. what would be an AI prompt to create customised reading materials and word lists?
ChatGPT: “Generate a 350-word short story for a 17-year-old student reading at an 1100 Lexile level with 8th grade vocabulary and 7th grade reading comprehension skills. The story should cover topics of high interest for 17-year-old students, use vocabulary appropriate for an 8th grade level, and have a 7th grade reading comprehension level in terms of sentence structure and complexity. Ensure accurate decoding of the text by focusing on phonetic regularity and avoiding exceptionally ambiguous letter combinations. Also generate a 150-word accompanying vocabulary list targeting 8th grade high utility academic vocabulary words that align with the content from the passage. Ensure the entire list can be decoded phonetically by the student. Provide definitions and use each vocabulary word in a contextualized example sentence. The materials should provide challenging yet accessible practice for the student's current reading abilities.”
This prompt provides key parameters to the AI around lexical complexity, topic, decoding level, vocabulary level, comprehension level, definitions and contextual word use. The AI could then instantly generate a custom passage and vocabulary list tailored specifically to William's needs. Teachers could review and tweak the output as needed before using it with Michael. But it provides a strong starting point unique to Michael versus more generalized materials.
Results
### Short Story: The Quest for the Solar Skateboard
Jason, a 17-year-old high school student, had always been passionate about two things: skateboarding and helping the environment. One sunny day, he discovered an article about solar-powered gadgets. An idea sparked in his mind. "What if I could build a solar-powered skateboard?" he thought. This skateboard would not only be cool but also eco-friendly.
With summer break ahead, Jason decided to embark on this ambitious project. He started by researching how solar panels work and how they could be integrated into a skateboard. With the help of his science teacher, Mr. Thompson, and his best friend, Mark, Jason began to gather materials: a skateboard, small solar panels, batteries, and a motor.
Working in his garage, Jason faced challenges. The biggest was figuring out how to make the solar panels strong enough to power the skateboard but light enough not to weigh it down. After many attempts, adjustments, and failures, they finally succeeded. The solar skateboard was not only functional but also fast.
The day of the big reveal arrived. Jason and Mark went to the local park, where friends and curious onlookers gathered. With a push, Jason hopped on the skateboard, and it smoothly accelerated. Cheers erupted as he maneuvered with ease, the solar panels glistening in the sunlight.
Jason's project was more than a success; it was a statement. He showed his peers that with innovation and determination, they could make a difference in the world, one solar panel at a time.
### Vocabulary List:
1. **Ambitious (adj.)** - Having a strong desire for success or achievement.
- Jason had an ambitious plan to build a solar-powered skateboard.
2. **Innovation (noun)** - The act of introducing something new or different.
- The solar skateboard was a perfect example of innovation in action.
3. **Eco-friendly (adj.)** - Not harmful to the environment.
- Jason wanted his skateboard to be not only cool but also eco-friendly.
4. **Maneuver (verb)** - To move skillfully or carefully.
- With a quick maneuver, Jason avoided the rock on the sidewalk.
5. **Glisten (verb)** - To shine with a soft, shimmering light.
- The solar panels glistened in the sunlight as Jason skated past.
6. **Determination (noun)** - Firmness of purpose; resolve.
- Despite the challenges, Jason's determination never wavered.
7. **Curious (adj.)** - Eager to know or learn something.
- The onlookers were curious about how the solar skateboard worked.
8. **Adjustment (noun)** - A small change made to something to improve it.
- Jason made several adjustments to the skateboard to perfect it.
9. **Functional (adj.)** - Designed to be practical and useful, rather than attractive.
- After many attempts, the skateboard was finally functional and fast.
10. **Gather (verb)** - To come together; to collect.
- Jason and Mark gathered all the necessary materials for their project.
11. **Spark (verb)** - To ignite; to trigger a reaction.
- Reading about solar-powered gadgets sparked Jason's imagination.
12. **Accelerate (verb)** - To increase in speed.
- The skateboard started to accelerate as Jason pushed off the ground.
Comments
The vocabulary list is designed to align with the content of the short story, offering high utility academic words for 8th grade level, ensuring phonetic regularity for accurate decoding. The whole exercise took less than 3 minutes to create - from the entering the description to get the prompt, to using the prompt to get the materials.
I share with parents that this is something they can do at home to further support their struggling students.
Give it a try and see what you think. For me, I don’t think AI will replace me as a teacher. But I do know that I love where AI is going in helping support me both personally and professionally.