“CATERing” to Readers’ Needs with AI: Innovation in Text Design and Instruction
I shared with Matt that we would be creating a book specifically for him. To write the book, I started by prompting AI to generate a six-sentence story about a soccer game with one-syllable CVC words that focus on short /i/ sounds and the first 100 Dolch words…. During a cold read of the text, Matt brought it to my attention that the images were more of a distraction than an added feature. He was excited, to say the least, to share ways of improving the book we made. After corrections, Matt focused more and we were able to engage with vocabulary and story structure. ChatGPT is a great resource to enhance your ideas, check for accuracy, and modify suggestions to suit unique needs.
—(Beverly, Matt’s mentor)
As the opening vignette suggests, in this Teaching and Learning in Action report, we describe the initial phase in a design/development study on supporting teachers in using Artificial Intelligence (AI) to write developmentally appropriate high-interest texts for students who need additional support in reading. Our broad goal is to enhance literacy teacher preparation by deepening teachers’ understanding of text features such as vocabulary load and decodability, their practices for choosing from existing texts, and their ability to adapt or write new texts to fit their students’ needs and interests. Toward this end, we have developed an innovation we call CATER (Computer-Assisted Texts for Early Reading) that expands teachers’ knowledge through the use of AI. CATER is designed for preservice and experienced teachers alike who aim to move beyond the constraints of the texts they have been provided by their school or district toward catering instructional texts to meet their students’ needs and interests.
During the summer of 2023, we piloted CATER in a graduate course. Our goals were to: 1. Develop a process for creating instructional books for children with AI; 2. Pilot
instruments to assess teacher learning and self-efficacy before and after the experience; and 3. Understand what benefits the books had for the children for whom they were designed. As we completed this pilot, we realized that the graduate students who participated in this research were teaching us things we had not anticipated about the potential for AI-supported text writing. Thus, they join us in this piece as co-authors and co-explorers of our new digital world underpinned by machine learning technologies. The aim of this article is to support beginning readers by sharing CATER and encouraging teachers to use AI to create books tailor-made for their students.
Take Action
- Identify a student who would benefit from having a text created just for them. This could be a reader who needs support with reading engagement and/or reading skills.
- Assess the students’ strengths and needs as areader with particular attention to their orthographic knowledge, vocabulary, and interests. Be sure to engage the student’s family/ caregivers in this process.
- Use AI to draft a text. Include: (A) words with orthographic structures the student is learning, (B) high-frequency
words, and (C) words in the student’s vocabulary. Be sure to position the student as the hero/ protagonist of the story. Adjust the AI-generated text as needed to match the student’s orthographic knowledge, vocabulary, and interests - Use AI to add images to the text. Decide whether you will share the text digitally or print, fold, and create a hard copy. Add images accordingly.
- Share the text with the student. Consider reading it chorally with the student and/or letting the student read it themself.
- Ask the student for feedback. Adjust the text again. Be sure to position the student as the context expert on the story. Send or digitally share a finalized copy of the text home with the student.