The Reading Crisis Nobody's Solving

The Evolution of EdTech — Part 11 of 13

By Clarence Stephen | May 2026

2,500 words | 10 min read

Here is a number that should keep you up at night: 130 million.

That's the number of American adults with limited literacy skills, according to the U.S. Department of Education and the National Center for Education Statistics (PIAAC Level 1 or below). One hundred and thirty million people. More than half the adult population of the richest, most technologically advanced nation in human history cannot comfortably read a newspaper editorial, a medication label, a lease agreement, or a children's novel.

If you're reading this article and following it without difficulty, you're in the minority. The majority of American adults would struggle with the text in front of you right now.

That fact alone should restructure every conversation about educational technology. It hasn't.

The Numbers

The reading crisis in America is not new. It's not a pandemic artifact, though the pandemic made it worse. It's not about smartphones or social media or any of the usual suspects we like to blame. It's a deep, structural, multi-generational failure that has been documented for decades and addressed with the urgency of a committee meeting about committee meetings.

The statistics are staggering. They're also consistent. And they're damning:

The economic cost, estimated by the National Literacy Institute: up to $2.2 trillion per year. That's roughly 10 percent of GDP — lost to reduced productivity, increased healthcare costs, higher incarceration rates, lower tax revenues, and the cascading frictions that pile up when more than half your population cannot comfortably process written information.

A Self-Perpetuating Crisis

Here's the thing about the literacy crisis that makes it so hard to solve: it reproduces itself. Mechanically. Across generations.

Children raised in low-literacy households have a 72 percent chance of growing up with low literacy themselves. A child whose parent can't read is unlikely to be read to. A child who isn't read to enters school already behind. A child who enters school behind stays behind — research on early reading gaps shows that children not reading proficiently by third grade are four times more likely to drop out of high school.

And the gap compounds along lines of race and income. Only 17 percent of Black fourth graders and 21 percent of Hispanic fourth graders scored proficient on the 2024 NAEP, compared to 40 percent of white fourth graders. That gap has barely narrowed in thirty years — a period spanning multiple reform movements, billions in funding, and the entire lifespan of the EdTech industry.

Where Is EdTech?

This is the question that haunts me: where is educational technology?

We have AI that can write poetry, generate photorealistic images, compose music, and pass the bar exam. We have adaptive learning systems that can teach calculus at midnight. We have gamification platforms that make vocabulary drills feel like video games. We have the most powerful information processing tools in human history.

And 130 million Americans can't read.

The EdTech industry — a market projected to reach $400 billion globally by 2030 — has, with a few honorable exceptions, largely ignored the reading crisis. The reasons aren't mysterious. They're economic.

The market prefers literate customers. Most EdTech products assume a baseline of reading ability that half the adult population doesn't have. Khan Academy's math exercises require reading word problems. Coursera's courses require reading lectures and assignments. Duolingo requires reading in your native language before learning another. If you can't read well, you can't use the tools designed to help you learn.

Reading intervention is hard. Teaching someone to read — or to read better — is fundamentally different from teaching them math or science. Reading isn't a single skill. It's a web of interlocking capabilities: phonemic awareness, phonics, fluency, vocabulary, comprehension, background knowledge, motivation, engagement. There's no equivalent of "2 + 2 = 4" in reading. Every text is different. Every reader is different. The interaction between reader and text is unique each time.

The ROI is unclear. Venture capital flows to products with measurable outcomes and short feedback loops. A math tutoring app can demonstrate test-score improvement within weeks. A reading intervention may take months or years to show results, and the outcomes resist simple quantification. Investors prefer the math app. The market, left to its own devices, solves the problems that are easiest to measure, not the ones that matter most.

Adults are invisible. The 130 million adults with low literacy are the most underserved population in American education, and they're nearly invisible to the EdTech industry. Adult literacy programs receive a fraction of K-12 funding. Technology designed for children is inappropriate for adults. And the stigma of adult illiteracy — the shame that prevents people from seeking help — creates a market that conventional channels can't reach.

What Current Solutions Look Like

The current landscape of reading technology is thin, underfunded, and not up to the scale of the problem.

Phonics and decoding tools — programs like Lexia and Reading Horizons — focus on the mechanics of reading: letter-sound correspondence, blending, word recognition. They're effective for beginning readers and backed by solid evidence. But they only address the foundation. A student who can decode words but can't comprehend their meaning isn't a reader yet.

Leveled reading programs provide texts at calibrated difficulty levels and track progress. Useful, but reductive — they treat reading as a linear climb through increasingly difficult texts, when in reality a student might read at a fifth-grade level on average but struggle with scientific texts and breeze through narrative fiction.

Audiobooks and text-to-speech are accessibility tools, not development tools. They make content available to people who can't read it, which matters. But they don't teach people to read. The distinction is the difference between a wheelchair ramp and physical therapy.

Comprehension programs — platforms focused on reading strategies like summarization, questioning, and inference — exist but are rare, limited in scope, and designed almost entirely for K-12 classrooms, not adult learners.

But here's where it gets real. What's almost entirely absent from the landscape is technology that addresses the deeper problem: engagement. The reason millions read below grade level isn't just a deficit of skills. It's a deficit of motivation. Reading, for many struggling readers, is an unpleasant experience — dense text, unfamiliar vocabulary, content that feels disconnected from their lives. There's no feedback, no reward, no reason to push through the discomfort. The reading just... stops.

The Visualization Gap

Here's something reading researchers have known for decades that the EdTech industry has largely ignored: comprehension is fundamentally a visual process.

When skilled readers read, they create mental images. They see the forest, feel the weather, picture the characters' faces. They construct what researchers call a "situation model" — a mental representation of the world described by the text — that enables inference, prediction, and retention.

Struggling readers often don't do this. They process text word by word, sentence by sentence, without building the mental model that transforms reading from decoding into experience. For them, reading isn't immersive. It's labor — translating marks on a page into sounds, without ever arriving at meaning.

Research by Paul van den Broek, Richard Mayer, and others has shown that visual representations — illustrations, diagrams, mental imagery strategies — significantly improve reading comprehension, especially for struggling readers. Allan Paivio's dual coding theory, proposed in 1971, holds that information processed through both verbal and visual channels is encoded more deeply and retrieved more easily than information processed through either channel alone. This has been replicated across hundreds of studies over five decades. It's among the most robust results in cognitive science.

And yet the vast majority of reading technology focuses exclusively on text.

Think about how absurd this is. We have AI systems that generate photorealistic images from text descriptions in seconds. We have models that can create visual representations of any scene, setting, or concept imaginable. And we have 130 million people who struggle to read in part because they can't visualize what they're reading.

The technology exists. The need exists. The connection hasn't been made.

Why This Matters

The reading crisis is not just an education problem. It's a democracy problem, an economic problem, a health problem, and a justice problem.

A person who can't read at a functional level can't navigate the healthcare system, understand a ballot, follow a medication schedule, help their children with homework, or participate fully in civic life. They're more likely to be unemployed, incarcerated, and in poor health. They're less likely to vote, to volunteer, or to trust institutions.

The $2.2 trillion annual cost isn't just lost productivity. It's lost human potential — 130 million people whose talents, creativity, and contributions are diminished because they struggle with the most fundamental skill of an information society.

Every other problem we're trying to solve — climate change, political polarization, economic inequality, public health — is harder to solve when more than half the population can't comfortably read the information needed to understand these problems. Literacy isn't one cause among many. It's the substrate on which every other form of engagement depends.

And yet the EdTech industry, flush with billions in venture capital and armed with the most powerful AI tools ever created, remains focused on teaching already-privileged students to code, helping college graduates earn professional certificates, and making vocabulary drills slightly more engaging for children who are already on track.

The 130 million are invisible.

A Different Approach

So. What would it look like to build technology that actually confronts the reading crisis?

It would need to meet readers where they are — not in a classroom, but on a phone, a tablet, a laptop. It would need to work with texts that readers want to read, not just those that educators prescribe. It would need to transform the reading experience itself — making it more visual, more interactive, more engaging — so that struggling readers have a reason to continue.

It would need AI that understands text well enough to visualize it. AI that adapts to individual reading levels. AI that can make any book, any article, any document more accessible without talking down to people.

It would need to be local, private, and affordable. It would need to serve a fourth-grader and an adult learner equally well. It would need to bridge the gap between text and meaning — between the words on the page and the world they describe.

In other words, it would need to answer a simple question: what if every story could be seen?

Someone is building that.

Next in the series: "What If Every Story Could Be Seen?" — the technology, the mission, and the live demo.

Sources

Clarence Stephen is the founder of Readify, an AI-powered literary platform. Yale B.Sc. Physics. Previously: Tiger Global, Morgan Stanley. Learn more at ireadifybooks.com

What could AI do for your classroom?

Readify uses local AI to adapt reading material to each student's level — generating illustrations, simplifying vocabulary, and tracking comprehension in real time. No student data leaves the building.

Bring Readify to Your School Try the Platform

Liked this article?

Get the full Evolution of EdTech series delivered to your inbox, plus weekly insights on AI in education.

Subscribe to Readify Branches