AI Is Causing ‘Cognitive Surrender.’ For Struggling Readers, the Risks Go Deeper

Last updated: June 29, 2026

Your child uses AI to look up a word, rephrase a sentence, or get unstuck on homework. It feels like a reasonable shortcut, maybe even a clever one. Now a U.S. Senate subcommittee has heard testimony giving that shortcut a name: cognitive surrender. Researchers at the University of Pennsylvania’s Wharton school define it as the moment AI stops assisting your thinking and starts replacing it, without you noticing the transfer happened. For most students that is worth watching. For a child still building the reading skills their brain has not yet wired in, the stakes are in a different category.

TL;DR

A June 16 Senate HELP subcommittee hearing on AI in K–12 schools heard testimony on “cognitive surrender” — Wharton research on AI replacing rather than assisting human thinking, without the person noticing.
84 percent of high school students are already using generative AI for schoolwork (Microsoft/Pew 2025–26); 85 percent of teachers use AI in their work.
Stanford’s SCALE initiative confirmed there are no high-quality causal studies yet on the long-term effects of AI on student learning.
For students whose phonological processing, working memory, and decoding skills are still developing, AI that bypasses those systems replaces the exact practice that builds reading pathways.
The useful distinction: cognitive offloading (freeing reasoning for harder work) vs. cognitive surrender (AI takes over the thinking entirely) — and the difference matters most where reading skills are still forming.

A Senate hearing named a new risk for students using AI: “cognitive surrender.” For children still building reading skills, here is what parents need to know.

Common questions

My child uses AI to help with reading. Is that a problem?

It depends on the use. Cognitive offloading — using AI to handle a peripheral task so your child can concentrate on harder thinking — is different from cognitive surrender, where AI replaces the skill your child most needs to practice. A text-to-speech tool that reads a history chapter aloud while your child focuses on the ideas is offloading; using AI to decode words or compose sentences for a child still building those skills is surrender. The key question: is AI freeing your child to think harder, or thinking for them on the skills their brain is still developing?

Doesn’t AI help kids with dyslexia by reading text aloud or fixing spelling?

Text-to-speech and spell-check can be valuable scaffolds when used alongside — not instead of — structured reading instruction. The research distinction matters: a scaffold that lets a child access content while still getting daily decoding and phonological practice supports learning. One that replaces that practice entirely leaves the underlying skill unbuilt. Brain-imaging studies show reading pathways form through practice; a child whose AI removes every decoding challenge may never build them. A screener is a starting point, not a diagnosis: if your child might need formal accommodations such as an IEP or 504 plan, or you suspect a vision, hearing, or medical cause, a professional evaluation is the route to those supports.

What is cognitive surrender, exactly?

Wharton Executive Education researchers define cognitive surrender as the point where AI is not just handling a specific task but making a decision — and the person adopts that decision as their own without recognizing the transfer happened. Using a calculator to do arithmetic is cognitive offloading: your reasoning stays in charge and moves on to harder work. Cognitive surrender is when AI provides the answer and you absorb it without the reasoning step. For students still building foundational skills, the transferred work is often the most important practice they are skipping.

What should I ask the school about AI use for my struggling reader?

Three questions worth asking: First, what specific AI uses are permitted or encouraged for students receiving literacy intervention, and is there an explicit boundary around decoding, phonological practice, and writing tasks? Second, how does the school track whether AI complements structured reading instruction or substitutes for it? Third — since Stanford’s SCALE initiative has confirmed there are no high-quality causal studies on AI’s long-term effects on learning — what is the default policy when evidence is thin for a child who is already behind?

Wharton researchers named it: 'cognitive surrender' means AI replaces thinking, not just offloads it. 84% of high schoolers already use AI for schoolwork. For kids still building reading skills, this is the wrong direction. Tweet Now Follow Learning Success

What the Senate heard

On June 16, 2026, the Senate HELP Subcommittee on Education and the American Family held a hearing on AI in K–12 schools. The numbers alone set the scene: 84 percent of high school students are already using generative AI for schoolwork, and 85 percent of teachers are deploying AI in their work, according to 2025–2026 research from Microsoft and Pew cited by Ranking Member Lisa Blunt Rochester.

Erin Mote, CEO of InnovateEDU and the EDSAFE AI Alliance, drew on Wharton Executive Education research to name the specific risk she sees in classrooms: cognitive surrender. She distinguished it from cognitive offloading — using a calculator frees your reasoning for harder work, so your thinking stays in charge. Cognitive surrender is different: AI produces the answer, and the person adopts it as their own without registering that the transfer occurred. The reasoning disappears, not just the calculation.

Delaware Secretary of Education Cynthia Marten framed the aspiration: “AI can expand opportunity while preserving what matters most about education — human relationships and human judgment.” Stanford’s SCALE initiative noted a sobering constraint: there are no high-quality causal studies yet on the long-term effects of AI on student learning. What the Senate called for, broadly, was teacher professional development and federal guardrails.

Author Quote "

AI can expand opportunity while preserving what matters most about education—human relationships and human judgment.

What the coverage gets wrong

Most reporting on the Senate hearing treats cognitive surrender as a uniform risk — a concern about students in general offloading their thinking to AI. The research on reading development makes the risk distribution uneven: students who are still building phonological processing, working memory, and decoding skills rely on practicing those systems to develop them. Brain-imaging evidence (Shaywitz, Yale; Temple, Stanford) shows that reading pathways form through intensive, appropriate instruction — meaning practice changes the brain. AI use that bypasses that practice does not affect all learners equally. For parents of struggling readers, the guardrails conversation is not abstract.

The risk is not evenly distributed

The hearing treated cognitive surrender as a concern for all students, which it is. What the coverage largely skipped is that the students with the least cognitive reserve for surrender are the ones who most need to build it. That matters most for struggling readers.

Reading is not absorbed the way spoken language is. Spoken language is biologically wired in; reading is a roughly 5,000-year-old cultural invention the brain repurposes from circuits built for other work, and it has to be taught, explicitly and systematically, every step of the way (cognitive scientist Stanislas Dehaene, David Geary’s primary-versus-secondary knowledge framework, National Reading Panel 2000). The International Dyslexia Association’s 2025 definition describes reading as a multi-system skill drawing on language processing, phonological awareness, working memory, attention, and processing speed simultaneously.

Now consider what happens when a child who struggles with phonological processing uses AI to decode an unfamiliar word instead of sounding it out. Or uses AI to rephrase a sentence instead of holding its meaning in working memory. That is not offloading a peripheral task; it is surrendering practice on the exact systems that need to be exercised. Brain-imaging studies from Yale (Shaywitz et al.) and Stanford (Temple et al.) showed that children who struggle to read develop the same reading pathways as typical readers after intensive, appropriate instruction: the brain physically rewires through practice. Neuroplasticity is real, which means practice matters, and the inverse is also true: routinely bypassing a struggling skill is a way to leave it undeveloped. The Senate is right that AI changes the cognitive landscape for every student. For students whose reading landscape is still being built, the change is sharper.

Key Takeaways:

Cognitive surrender: Wharton researchers define it as AI making a decision that a person adopts as their own without recognizing the transfer — distinct from useful cognitive offloading.

Why struggling readers are most exposed: Reading pathways form through practice; Yale and Stanford brain imaging shows intervention physically rewires the brain. AI that bypasses phonological and decoding practice leaves those pathways unbuilt.

The question to ask: Is AI freeing your child to think harder about something, or is it doing the cognitive work your child most needs to practice? The distinction determines whether it helps.

What this means at home

None of this is an argument against AI tools in education. Some uses are genuine offloading — a text-to-speech tool that reads a history textbook aloud so a child can absorb the content while their decoding skills are still catching up is a scaffold, not a surrender, as long as the child is also getting structured reading instruction elsewhere. The distinction Wharton researchers draw is useful for parents: is the AI freeing your child to think harder about something, or is it thinking for them on the skills they most need to practice?

Three questions worth asking the school: First, what specific AI uses are permitted or encouraged for students receiving literacy intervention, and is there an explicit boundary around decoding, phonological practice, and writing tasks? Second, how is the school tracking whether AI use is complementing structured reading instruction or substituting for it? Third, as Stanford’s SCALE noted, the long-term causal evidence on AI and learning does not yet exist — what is the school’s default when evidence is thin for a child who is already behind?

The Senate hearing called for protecting human judgment in schools. For a child still building the cognitive architecture that skilled reading requires, the most important human judgment in the room is yours.

Reading is built, not downloaded. The brain that struggles to read today is not the brain your child will have after a year of the right kind of practice — that is what the imaging actually shows, not a motivational poster. The obstacle is not AI. It is the quiet assumption that a tool producing the right output is the same as a child developing the skill to produce it. Those are not the same thing, and the difference compounds over time. If you want to know exactly which cognitive systems your child is still building and a plan to strengthen each one, the All Access program gives you that roadmap, with a free trial and a personalized Action Plan you keep regardless.