When I was deputy chancellor, I visited a 1st grade classroom to see how early-literacy practices were taking shape. The teacher invited me to sit beside a small group as they practiced with decodable readers. As students sounded out “ship,” “shop,” and “shut,” an iPad at the table captured brief oral-reading clips. After class, the teacher and her coach opened those recordings to listen for error patterns, confusions with /sh/ vs. /ch/, and missed short-u sounds buried among look-alike words. She was gathering information that necessarily influenced tomorrow’s mini-lesson. That is what AI in literacy should feel like: more informed and personalized instruction, not less human teaching.
AI can work best in teaching literacy when it is used in service of the science of reading—phonemic awareness, phonics, fluency, vocabulary, and comprehension—and delivered by well-prepared educators with the time and support to do the work effectively.
What Research Really Tells Us
A new national survey of 468 elementary teachers found that nearly every teacher (99.6 percent) used digital texts during the 2020–21 school year, and most planned to keep using them the following year (Tortorelli & Strong, 2025). But the study also surfaced a familiar pattern: Teachers often used digital texts as substitutes for print and defaulted to read-aloud mode even when their goal was to build decoding or fluency. Only about 29 percent reported receiving professional learning on how digital tools support literacy development, and just 17 percent received resources aligned to instructional goals. In short, access to technology grew faster than the technological–pedagogical–content knowledge (TPACK) teachers needed.
When researchers look across dozens of studies comparing digital and print reading, two clear takeaways emerge. First, students, especially younger ones, often understand informational texts better in print unless the digital version is designed to match what good instruction looks like. For example, adding animations or sound effects that don’t connect to meaning can actually distract from learning (Clinton, 2019; Furenes et al., 2021).
Second, when digital texts include high-quality, supportive features, they can help students grow in vocabulary and comprehension. Features like interactive dictionaries, embedded questions, or short prompts that nudge readers to think more deeply about what they’re reading make a meaningful difference (Savva et al., 2022). And for our youngest readers, word-level features that say a word aloud, highlight letter–sound patterns, or segment words into sounds can strengthen foundational reading skills when teachers use them with purpose (Korat & Blau, 2010).
But for teachers to be able to use new technologies with purpose, schools and districts must invest in professional learning. Teachers need structured opportunities to connect research on how students learn to read with the tools now shaping literacy instruction, including how to tell when digital texts enhance learning and when they simply add noise.
Ultimately, the goal of reading instruction hasn’t changed in the AI era. What has changed is our capacity to see and respond to learning in new ways. When AI strengthens the feedback loop between teachers and students, illuminates where a learner is stuck, surfaces patterns in oral reading, and prompts reflection, it becomes a partner in learning, not a shortcut.
AI Can Help, If We Stay Anchored to the Science
For teachers and coaches, edtech tools show real promise when they stay grounded in what we know about how children learn to read. The following two examples show how we can thoughtfully pair emerging technology with human-to-human support and other forms of structured guidance to yield the best results for teachers and, ultimately, students.
1. Low-Inference, High-Frequency Feedback
New AI tools can help teachers see patterns they already look for (who’s doing the talking, the kinds of questions being asked, which phoneme–grapheme patterns trip readers up) without waiting weeks for a coaching visit. At Teaching Matters, a nonprofit professional development organization for educators, we piloted an AI-enabled classroom talk tool in secondary math to strengthen questioning and student discourse. The tool securely records a class session and instantly analyzes the talk patterns to discern how much students are talking, the levels and types of questions teachers ask, and whether discussion opens space for deeper thinking. It doesn’t evaluate instruction; it simply listens to the flow of the lesson and converts it into low-inference data teachers and coaches can use right away.
We’re now studying how similar fast-feedback cycles can support literacy coaching so teachers can adjust between visits, not just after them. The goal isn’t evaluation—it’s timely insight that helps teachers refine instruction in the moment.
For motivation, the most effective digital design element isn’t a game, it’s offering relevance and choice.
2. Just-In-Time Supports for Foundational Skills
Early reading platforms that “listen” as children read aloud and offer targeted, immediate feedback mirror the kind of practice I described earlier with the teacher who captured brief oral reading clips on an iPad and used them with her coach to plan tomorrow’s mini-lesson. Tools like these can support that same cycle of quick insight and responsive instruction and can make a difference when teachers use the data to inform small-group instruction and lesson design. Reporting on the AI tool Amira, for example, The Hechinger Report (Mader, 2021) described schools seeing double-digit gains in oral reading fluency and reminded readers that human judgment still matters, especially for multilingual learners, dialect variation, and speech differences.
Our own early pilots at Teaching Matters told a similar story: When students used adaptive reading tools on their own, results were uneven. But when those same tools were paired with coaching, structured routines, and guided reflection, growth accelerated. Technology didn’t drive improvement; the human interactions that generated high-quality teaching did.
At the national level, the Institute of Education Sciences launched the Center for Early Literacy and Responsible AI, which is developing an “AI Reading Enhancer” to generate personalized decodable texts, analyze oral reading in real time, and deliver just-in-time supports for culturally and linguistically diverse learners, all with an emphasis on ethical, transparent use (Institute of Education Sciences, 2024).
I had the opportunity to hear university researchers John Tortorelli and Laura Strong share findings from their national survey of elementary teachers at the Amplify & Elevate Innovation: AI in Education Symposium at the University at Albany in July 2025. Their research shows how quickly access to digital texts has outpaced the technological/pedagogical/content knowledge teachers need to use them well. But, it also points to a hopeful shift: The field is moving beyond “more devices” toward smarter, research-aligned use of technology, where AI deepens teacher insight and strengthens the science of reading in practice.
A Feature-to-Skill Playbook
If you’re a teacher or coach trying to guide struggling students to read, try starting with this simple question before opening your laptop: What skill am I teaching and which digital feature actually supports it?
For phonemic awareness and phonics (K–2), tools that allow word-level playback, sound-by-sound segmenting, or letter–sound highlighting can be powerful allies (Korat & Blau, 2010). Navigating decodable texts using spelling patterns keeps practice aligned to instruction.
What’s less helpful is defaulting to full-text read-aloud mode because it removes the very decoding work students need to practice. When word-level supports are used intentionally, however, they can boost early phonological awareness and word reading (Chera & Wood, 2003).
For fluency (grades 1–3), prioritize tools that let students record and review their oral reading, highlight text as they reread, and time rereads with a clear purpose to improve phrasing, expression, or pacing. The evidence for “read-aloud to build fluency” is limited; fluency develops most when students do the reading and rereading and receive targeted feedback in real time (Swanson et al., 2020; Tortorelli & Strong, 2025).
For vocabulary and comprehension (K–5), digital tools can truly shine when they support active thinking. Interactive dictionaries, embedded questions with immediate feedback, and annotation tools that invite students to capture their ideas all help make meaning visible. Digital enhancements that deepen understanding rather than distract from it, such as a hotspot that lets students tap a word to see a kid-friendly definition, image, and example sentence, or a prompt that pauses the text and asks students to predict what will happen next based on the evidence, can strengthen vocabulary and comprehension more than static digital or paper versions (Savva et al., 2022).
This same principle extends into the upper elementary grades. Rather than withholding complex texts until students master foundational skills, approaches like the Read STOP Write strategy embed decoding, fluency, and metacognitive supports directly into students’ work with grade-level texts, strengthening both comprehension and writing (Strong et al., 2025).
And for motivation, the most effective digital design element isn’t a game, it’s offering relevance and choice. Students engage more deeply when they see themselves in the texts they read and when topics connect to their interests and experiences. Extrinsic motivators like badges and rewards may catch attention, but they rarely build sustained engagement or skill, especially for struggling readers (McGeown et al., 2016).
In the end, AI and digital tools strengthen literacy instruction by helping teachers make the right move at the right moment, not by replacing the work of learning or doing the thinking for students.
Four Moves for Schools and Districts
Taken together, the following four moves offer a practical roadmap for schools and districts seeking to align digital and AI-enabled tools with evidence-based literacy instruction.
1. Replace “press play” with “feature-fit.” When visiting classrooms or leading PD, leaders and coaches should shift the conversation from whether teachers are using digital texts to how they’re using them. Ask: Which feature and why? If the goal is decoding, you should see students using word-level supports, and segmenting tools, not just listening passively to full-text read-alouds. The goal isn’t more screen time; it’s more intentional use of features that actually build reading skills.
2. Build professional learning around TPACK for early literacy. Short, focused sessions (45 minutes is plenty!) can make a big difference. Try framing professional learning around practical questions like, “When should I avoid read-aloud mode?” or “How can digital segmenting tools reinforce today’s phonics pattern?” Teachers in Tortorelli and Strong’s (2025) national survey didn’t lack willingness to use digital texts; they lacked the aligned training to use them well.
3. Insist on alignment with high-quality instructional materials. Choose digital tools that strengthen—not compete with—your high-quality core literacy materials. The best platforms surface instructionally relevant data, not distractions. If a feature pulls attention away from text or comprehension, skip it. Look for interoperability and, more importantly, ensure alignment and clear connections across all tiers of instruction.
4. Lead with ethics, safety, and guidance. Any AI-enabled system must protect student and teacher data through secure, district-approved platforms. Beyond compliance, model transparency: Teachers and families should always know what data are collected and how they’re used. Build clear guardrails so AI enhances instruction rather than evaluates it.
No Shortcuts Here
I often say that this work—the leadership and instructional decisions that shape how students experience learning—isn’t about making schools more efficient; it’s about deepening student thinking. In literacy, that begins with sound-symbol mapping and grows into knowledge-rich comprehension, and it runs on the steady craft of teachers. AI can help us see more, sooner. But the meaningful work, the noticing, prompting, reteaching, and celebrating, still belongs to us. Let’s use edtech tools to strengthen the science of reading, not shortcut it.
Reflect & Discuss
Think about a recent lesson where you used a digital tool to help students with their reading. What features of the tool worked well, and which seemed distracting or unnecessary?
What’s the difference between a digital tool that “supports” the science of reading versus one that simply “delivers” content digitally? Why does this distinction matter?
