This article was written and published by Jennifer Vilcarino — July 3rd, 2025
What barriers do students face when asked to read text and show their comprehension of it?
This question was posed by an English/language arts teacher and educational consultant to a group of educators at a session at the ISTELive 25 + ASCD Annual Conference 25 here, held June 29 to July 2.
Some of the common responses among the audience of mostly teachers included limited vocabulary, boredom, and difficulty decoding.
The session—led by Jessica Pack, a 6th grade language arts teacher at James Workman Middle School in Riverside County, Calif., made the case that smart, strategic use of artificial intelligence tools could help boost reading skills. (However, it’s important to note that many educators say AI tools do the exact opposite—they stifle creativity, lead to plagiarism, and give students an easy way out of tackling challenging assignments.)
Improving reading skills is one of the top priorities in schools across the country because of data showing that those skills are declining. The 2024 National Assessment of Educational Progress has shown declining reading scores for 4th and 8th graders for years.
Those trends have prompted growing interest in the “science of reading,” the use of proven, evidence-based methods for teaching reading that include teaching the foundations of language in a structured progression.
Pack has been using AI tools to bolster reading comprehension lessons, she said during the ISTE session, titled “Enhancing Literacy Through Creativity Using AI-Powered Tools.”
“A lot of folks are landing on AI as a purely teacher-centered type of tool, so what we are going to do today is encourage a bit of student-centered use,” Pack told the audience of educators at the conference here.
How this teacher used an AI image generator to work on reading skills
Pack said her 6th grade students from last school year had an average reading level of about 2nd or 3rd grade.
To address this learning gap, students were first asked to examine paragraphs from a book and generate keywords from that text. They would use those keywords to create a prompt for an AI image generator.
The students would then work in small groups to evaluate the image to see if it was missing something that was important to the text from the book. Finally, they would submit additional AI prompts to get the image to more accurately represent the text.
“This is huge. This is the metacognitive moment where they’re demonstrating their comprehension—their full comprehension of the text—by being able to connect it back to whatever it is they were able to generate,” said Pack.
Research shows that asking students to monitor and correct their own understanding of text as they read can boost their comprehension—though using AI for this purpose hasn’t been studied extensively.
Pack emphasized that it is important for students to be taught to cite the images they create as generative AI images, because that instills the value of citing sources for content they create.
“They need to be aware of what content is AI-generated, and it starts building that digital citizenship foundation for citation of AI and being aware that not all the things we are creating are actual factual things,” she said.