How Technology Can Help Close the K–12 Literacy Gap — If Schools Use It Well

Reading technology has become much more capable over the past few years. Schools can now use digital screeners to spot recurring errors, adaptive platforms to adjust practice on the fly, and newer AI tools to listen to students read aloud and flag patterns that a teacher may not catch in one short conference.

That sounds impressive, but I am less interested in what the software can technically do than in what happens five minutes later.

EdTech Magazine recently reported on districts using technology to strengthen literacy instruction as reading performance remains a concern across K–12. What caught my attention was not the usual promise of “personalization”, schools have been hearing that word for years. The more useful development is that reading issues previously lumped together under the general category of dyslexia can now be differentiated using these techniques, allowing teachers to target specific areas for improvement in students’ reading abilities.

Take two fifth graders who land in roughly the same low band on a benchmark assessment. On paper, they look similar, but in class, they may be nothing alike. One student gets stuck on multisyllabic words and starts decoding almost letter by letter, another reads the passage with confidence and expression, but both of them go blank when asked why the main character changed course halfway through. I have always thought this is where broad reading scores become frustrating: they tell me who is struggling, but not always what the struggle actually is, so I have no idea how to help them. And there are quieter cases too. A student may read smoothly until the text fills up with words such as reluctant, consequence, or subsequent. At that point, the problem is no longer fluency, it is vocabulary, background knowledge, or both.

None of these classroom responses is revolutionary. Experienced teachers have been making these distinctions for years, based on them, they can provide personalized adjustments and guidance to their students. In my opinion, what software can offer is speed, instead of spending three weeks comparing worksheets, notes, and benchmark results, a teacher may see the pattern much sooner.

Data Is Cheap. Useful Signal Is Not.

Schools already collect a ridiculous amount of information: attendance, reading scores, benchmark data, standardized tests, classroom assignments, intervention logs, behavior records… Then each platform produces its own dashboard, which is supposed to make everything clearer and sometimes manages to do the opposite. Because simply raising test scores does not allow teachers to gain a comprehensive understanding of the students’ situation or devise appropriate strategies.

If a student scores 68 percent on a reading assessment, I still do not know very much, but if the same system shows that the student gets literal-recall questions right but repeatedly misses inference questions, now there is something a teacher can work with. The next lesson may involve stopping after each paragraph, asking what the author is implying, and requiring the student to point to evidence rather than simply retell the text. Now imagine another student with falling reading scores who has also missed 14 days of school. That is a different problem. I am wary of education technology that treats every weakness as something to be fixed inside the platform. Sometimes the correct response is a reading intervention. Sometimes the student needs to be back in class consistently before another piece of software will make much difference.

Personalization Gets Overhyped

“Personalized learning” is one of those phrases that has been stretched so far that it can mean almost anything. For literacy, I prefer a much narrower version: give the student work that actually matches the problem.

A child who cannot decode unpredictable needs something different from a child who can pronounce the word but has no idea what it means in context; A fluent reader who cannot follow an argument through a long passage does not need more phonics practice. Yet I still see products marketed as though simply changing difficulty level counts as personalization.

Sometimes it does. Often it does not.

The better systems can help teachers sort students more efficiently, especially when one teacher is trying to manage 25 or 30 readers at very different levels. That has real value. But the software only sees what the student did on the task in front of it, it does not know that the child slept badly the night before, recently moved schools, is learning English, or misunderstood the topic rather than the reading skill.

A score looks clean, but children rarely are.

That is the part I would not hand over to an algorithm.

AI Reading Tutors: Useful, Up to a Point

AI-powered reading tools are where the conversation gets more interesting.

Some can now listen to oral reading, detect recurring pronunciation errors, track fluency, and respond almost immediately, in a busy classroom, which can help. A teacher simply does not have time to sit beside every child for ten uninterrupted minutes each day. Here is the use case that makes sense to me: a tool notices that several students consistently stumble over the same vowel pattern, the teacher sees that pattern, pulls those students together the next morning, and spends ten focused minutes addressing it.

Good.

But what I do not want is the diagnosis becoming the lesson.

A student misreads five words, the platform generates five more activities. The child clicks through them, gets a green check mark, and moves on. Nobody asks whether the student can now recognize the pattern in a new text. That is the sort of thing that makes classrooms look technologically sophisticated while the instruction underneath barely changes.

Frankly, there are plenty of products in education that seem to exist because “AI-powered” looks good in a sales deck.

My patience for those tools is limited.

If the software helps a teacher notice something useful, I am interested, but if it mainly gives students another screen, another login, and another progress bar, I am much less impressed.

Reading Recovery Is Still Mostly About Boring Things

This is probably the least exciting part of the story, but it matters most: Children get better at reading by reading.

They need explicit instruction when decoding is weak, they need vocabulary; they need to discuss texts, reread confusing passages, make mistakes, hear stronger readers, and spend enough time with language for patterns to become familiar.

They also need to show up.

Chronic absenteeism does not become less damaging because a district has good software, but reading skills accumulate over time, and missed instruction compounds in ways that are easy to underestimate. EdTech has covered districts using technology to identify attendance problems earlier and communicate with families more quickly. That kind of system can help schools act sooner, which matters for literacy, even though the tool itself does not teach reading.

This is also why I would be cautious about attributing too much literacy improvement to software alone. A district may adopt a reading platform at the same time it expands tutoring, improves attendance follow-up, changes curriculum, or gives teachers better training, real classrooms are messy like that, the clean cause-and-effect stories usually belong in marketing materials.

What I Would Keep

With pandemic-era funding gone and school budgets under more pressure, districts are being forced to decide which platforms are actually worth paying for. I think that is healthy. For several years, schools accumulated tools very quickly, some were useful, some were redundant, some were purchased because there was funding available and a problem that needed to be solved immediately.

Now the standard should be tougher.

When evaluating a literacy product, I would skip most of the vendor language about machine learning, adaptive pathways, and intelligent personalization, I would ask the teacher one question:

Does this tell you what to do differently tomorrow morning?

If a dashboard shows that a child is struggling but leaves the teacher guessing about the next instructional step, I do not see much value in it. If it helps distinguish decoding from vocabulary, fluency from comprehension, or weak reading from an attendance problem, then it may be doing something worthwhile.

That is where I think technology earns its place in literacy instruction, not by becoming the lesson, and certainly not by replacing the teacher’s judgment, but by clearing away enough noise that the teacher can see the student more clearly.

For me, that is a much more convincing promise than “AI-powered reading.”

Simon Holmes | Contributing Education Reporter

Simon Holmes is an education reporter who contributes to WuKong Education, covering K–12 education, education technology and issues affecting students and families around the world.

With a background in education and experience as a teacher, Simon brings firsthand insight into classroom learning, student development, and the challenges educators face.

For WuKong Education, he reports on education research, emerging learning trends, policy developments, technology in the classroom, and issues relevant to parents, teachers, and students.