Using Technology to Support Students’ Understanding of Nonfiction Reading

Instead of placing restrictions on the use of technology in the classroom, what if we were to promote its responsible and productive use to help students better understand nonfiction reading? We propose that search engines like Google or generative AI technology like ChatGPT can be leveraged to make expository text more accessible, make timely implantation of freshly reported science news easier by creating leveled texts, and help bridge gaps between separate disciplines.

Using AI to Level Text and Identify Vocabulary

We—science teacher Shawn and English language arts teacher Soojung—collaborated on a DNA unit that took us about two weeks to plan. We decided to use ChatGPT to determine if our selected text was at an eighth-grade level.

We prompted ChatGPT to identify a text’s level by using conversational directives: “Identify the grade reading level of the following text: [copy and paste article].” 

ChatGPT gave varied answers, even with the same prompt. Sometimes, it provided the reading level alongside an explanation under four subheadings: (1) sentence structure, (2) vocabulary, (3) conceptual depth, and (4) clarity and accessibility. Other times, ChatGPT provided a rationale based on methodologies such as the Automated Readability Index and others. 

Once ChatGPT provided us with the reading level of our text, we leveled the text by prompting ChatGPT: “Level the following text at the eighth-grade reading level: [copy and paste article].” This allowed all students to participate with essentially the same text but leveled for individualized needs. 

Furthermore, we prompted ChatGPT to identify key vocabulary: “From the following text, identify critical vocabulary for eighth graders to preview/understand: [copy and paste article].” 

Visualizing Vocabulary With Search Engines

When students read articles with complex terminology, they may lose sight of the fact that the terms represent concrete objects or observable processes. Shawn guided students as they deciphered the leveled article on DNA structure: “I want you to be able to picture what you are reading—especially the difficult words.” 

We used Google’s Advanced Image Search to find visual representations (from ChatGPT) of challenging DNA vocabulary. Teacher guidance was necessary to help students identify the best visual representation of the term, as many of the images in the search results were above middle school level. Once the class agreed on an accurate model for the term, we illustrated that vocabulary in the margins of the article. This form of annotation allowed students to see connections between structures of the DNA molecule, like “nucleotides” and “nitrogenous bases.”

Using Search Engines and AI to Clarify Terms

Students also used the assistance of technology to improve their comprehension of science-specific text. The descriptive phrasing and vocabulary can be a stumbling block for students who aren’t familiar with technical nonfiction writing. 

If students encountered challenging words or phrases, they were encouraged to Google synonyms of those words to improve their understanding. A student annotated the term sequence, for example, and retrieved simpler synonyms, such as order, series, or string.

Students can utilize AI as a more responsive tool than a search engine. If a learner encounters a difficult passage, the student can prompt ChatGPT to rewrite the sentence at a lower reading level so that it is easier for them to understand: “Rewrite the following sentence at a sixth-grade reading level…”

Using AI to Enhance On-Topic Multimedia

To help solidify student understanding of DNA, we selected Judith Hauck’s video The Twisting Tale of DNA, posted on TED-Ed. By collaborating with another content teacher, students experience knowledge as interdisciplinary.

Vizard.AI, a generative AI tool, allows educators to quickly create a transcript and/or subtitles to accompany video so that students can experience domain-specific vocabulary, such as “Escherichia coli,” in an audiovisual context. This reinforces their vocabulary acquisition through audiovisual cues (imagery, pronunciation, and subtitles). 

In Soojung’s English class, students identified concepts within the video that they had previously encountered in science. Students observed how prior knowledge of DNA from science class alongside multimedia elements, like animation, led to a heightened understanding of DNA and helped them grasp the content of the video. Additionally, students explored how multimedia elements complemented nonfiction concepts, like how narration, animation, and subtitles enhanced their understanding of DNA. 

Enhancing Science Vocabulary with Figurative Language

“Modeling” is a science and engineering practice that is emphasized in the Next Generation Science Standards. Models aren’t limited to physical representations. Once students grasp new vocabulary with the help of AI, they can model connections between concepts using literary language. Figurative language, like analogies, can be a familiar, powerful way for students to make abstract ideas more accessible. 

Learners make the connection of the DNA molecule being like blueprints that a home builder uses, or how nitrogenous bases are found in fixed, predictable partnerships, similar to best friends who are always found together. A practical requirement for student analogies is that a nonscientist would be able to read them and clearly understand the connections.

Modeling Ethical AI Usage

Nothing trumps teacher expertise. OpenAI, the organization that developed ChatGPT, includes a disclaimer below the chatbox: “ChatGPT can make mistakes. Check important info.” Proofread its output for errors and misinterpretations. Communicate this best practice with students. 

When distributing works to students, make sure to credit the original author and note the following: “..., adapted by [your name] and OpenAI’s ChatGPT (Version 3.5).” OpenAI’s “Sharing and Publication Policy” allows free sharing of content created using their software; however, the content must be identified as coauthored by AI. The user shoulders all accountability of shared content.

Invite students to notice AI attributed as a coauthor. Consider expanding the conversation to transparent ethical usage of generative AI and communicating how different educators and institutions have varying policies and acceptable usage of AI.