What Can AI Chatbots Teach Us About How Humans Learn?

The lowest-performing middle school readers aren't just behind, they're falling further behind every year, even as reading scores recover elsewhere.On This Week with EdSurge, NWEA's Megan Kuhfeld shares new research that explains why the K-3 literacy laws that fixed early reading aren't reaching the middle school.Then Dacia Toll, who spent years running schools before co-founding the AI literacy platform Coursemojo, describes what she calls the comprehension wall, the point where a middle schooler can read every word on a page and still miss the meaning, and makes the case for AI tools built to ask questions instead of just answering them.WHAT YOU'LL LEARNMegan Kuhfeld's NWEA research shows the same decline at the 10th percentile of middle school reading scores across three separate assessments: NWEA's MAP Growth, i-Ready and NAEP, drawing on data from more than 20 million K-8 students.Dacia Toll's "background knowledge problem" means two students with identical reading skills can perform two full grade levels apart on the same text, depending on how much they already know about the topic.K-3 literacy laws mostly stop at third grade and many didn't take effect until 2022-2024, so they miss students who are already behind by the time they reach middle school.An individual struggling reader typically needs two to three years of above-average growth to catch up to grade level, and building a system that catches every student takes even longer.There's a meaningful difference between "indulgebots," AI tools that just answer a student's question, and purpose-built instructional AI that responds to a wrong answer by sending the student back to a specific sentence or paragraph to work it out themselves.STORIES MENTIONED IN THIS EPISODEReading Scores Continue to Drop for Many Middle Schoolers by David WeldonAI Can Help One of K-12's Biggest Challenges: Middle School Reading by Dacia TollSTAY CONNECTEDSubscribe to EdSurge newsletters at edsurge.com.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing:Megan Kuhfeld, PhD, Director of Growth Modeling and Data Analytics, NWEADacia Toll, Co-Founder and Co-CEO, Coursemojo

Two Voices of Change Fellows had the same kind of moment this year, when a student comment that could have been brushed off as attitude turned out to be a warning sign. On This Week with EdSurge, Melinda Medina explains what happened when a student answered a literature question with "I use ChatGPT, how am I supposed to know that?" and what it revealed about students quietly outsourcing their own thinking. Then Beth Yirga describes the eighth grader who asked her, "If AI will do everything, then what's the point of learning?" and how that question sent her looking for answers outside of education entirely.WHAT YOU'LL LEARNMelinda Medina cites a Brown University economics class where students averaged nearly 96% on a take-home midterm, then under 49% once the professor switched to an in-person proctored exam, a trend she now sees echoed in her own classroom.When Melinda's 11th graders were asked how author Kate Chopin uses irony, one replied that she'd used ChatGPT instead of working through the text, a moment Melinda says worried her less as cheating and more as a sign students no longer wanted to engage with productive struggle.Beth Yirga's question came from an eighth grader during a planning meeting for a new career exploration program at Freire Charter School Wilmington, and it sent her to the Women Who Code Summit in New York, where she was the only K-12 educator in a room of more than 500 people.Beth is building an "identity year" for students, pairing AI literacy with outdoor hikes, a trip to the opera, and a visit to the state capitol, on the belief that students need to know who they are before they can direct any tool toward a future worth having.STORIES MENTIONED IN THIS EPISODEProtecting Student Cognition in the Age of AI by Melinda Medina The K-12 Silo Won't Survive AI, But Our Relationships Might by Beth YirgaSTAY CONNECTEDSubscribe to EdSurge newsletters at edsurge.com.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing:Melinda Medina, English and Special Education Teacher, Brooklyn, NYBeth Yirga, Assistant Head of School, Freire Charter School Wilmington, DE

New guidance was supposed to help teachers make sense of screen time. New laws were supposed to keep phones out of the way. On This Week with EdSurge we look at what happens when teachers are left to make sense of rules that were never built with their classrooms in mind. Adam Stone talks with pre-K and kindergarten teachers about the guidance that excludes them, and civics teacher Dan Clark explains why he thinks Missouri got its phone ban wrong.WHAT YOU'LL LEARNThe American Academy of Pediatrics updated its screen time guidance for the first time in a decade, but the update is scoped around children and adolescents and offers little direction for pre-K classrooms.Pre-K and kindergarten teachers are relying on tactile, low-tech strategies, such as forming letters out of Play-Doh, to reinforce face-to-face interaction instead of defaulting to screens.Missouri's new law bans any portable communication device on a student's person from the first bell to the last, which cut off students working on a community advocacy project from the platform they knew best how to use.We explore if blanket phone bans skip the harder work of teaching digital citizenship, comparing the debate to longstanding arguments over whether schools should still teach cursive.STORIES MENTIONED IN THIS EPISODEEarly Childhood Education Teachers Grapple with Screen Time by Adam StoneThe Case Against the Classroom Cellphone Crackdown by Dan ClarkSTAY CONNECTEDSubscribe to EdSurge newsletters at edsurge.com.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing:Adam Stone, Education and Technology JournalistDan Clark, Eighth Grade Civics Teacher, Missouri

The Department of Education just issued its first major classroom technology guidance in years, urging schools to judge tools by outcomes rather than screen time. Digital Promise president and CEO Jean-Claude Brizard and outcomes-based contracting expert Brittany Miller join This Week with EdSurge to break down what the new Dear Colleague Letter gets right, what it leaves out, and what happens now that Washington has handed implementation to the states. They agree the letter marks real progress, but they do not agree on how much it accomplishes.WHAT YOU'LL LEARNThe Dear Colleague Letter draws a clear line between recreational and instructional technology use, giving school leaders new language to push back against blanket bans, but it also leans on randomized controlled trials as a measure of evidence, a standard Brittany Miller says moves too slowly for how quickly edtech itself changes.Ten states already run statewide outcomes based contracting efforts, including a new Michigan law tying fifty million dollars in tutoring funding to results, which Miller points to as proof the model is working even without a federal mandate behind it.Jean-Claude Brizard argues schools should demand third party certification rather than commission their own research, compares scattered anti-technology sentiment to the anti-vaccine movement, and calls on school and district leaders to reclaim the public narrative around edtech.Both guests agree the guidance leaves out direct support for the teachers, principals, and superintendents actually implementing it, and both point to procurement, not classroom policy, as the real lever for turning the letter into practice.STORIES MENTIONED IN THIS EPISODEDepartment of Education Issues Long-Awaited Edtech Guidance for States and Districtsby EdSurge StaffSTAY CONNECTEDSubscribe to EdSurge newsletters at edsurge.com/newsletters.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing:Jean-Claude Brizard, President and CEO, Digital PromiseBrittany Miller, Center for Outcomes-Based Contracting

This week, two guests trace the true cost of AI, one in the classroom and one behind the scenes. Michael Hernandez explains how he taught himself to code so he could build a free tool that flags weak assignments before students ever think about cheating. Mi Aniefuna follows the electricity powering that same technology, tracing the invoice that lands on a district's desk long after the contract is signed. Together, their stories ask what schools are really paying for when they bring AI into the building.WHAT YOU'LL LEARNMichael Hernandez used vibe coding, describing an idea in plain language and letting AI write the code, to build the Cheat Vulnerability Index, a free tool that scans an assignment and flags where it is exposed to cheating.Michael identifies three traits, originality, personal connection, and purpose, that he believes make an assignment resistant to cheating.Mi Aniefuna reports that United States data centers used about 176 terawatt hours of electricity in 2023, roughly 4.4% of the countrys total electricity consumption.Mi points to a Stanford review of more than 800 studies on AI in K-12 education, only 20 of which met the bar for rigorous, causal-impact research, even as adoption keeps outpacing the evidence.STORIES MENTIONED IN THIS EPISODEStory 1 I Used AI to Build AI-Resistant Assignments by Michael HernandezStory 2 Can Schools Afford an AI-First Future? by Mi AniefunaSTAY CONNECTEDSubscribe to EdSurge newsletters at www.edsurge.com/newsletters.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing:Michael Hernandez, Educator and AuthorMi Aniefuna, Public Health and Behavioral Sciences Researcher, EdSurge

What does it actually take to support the adults who shape our youngest learners? Nicol Russell, chief academic officer at Teaching Strategies, argues that early educators deserve the same intentional investment as engineers, not treatment as replaceable hires. Dr. Nneka McGee, founder of Muon Global, brings that same philosophy to how five year olds encounter AI, starting with conversation instead of screens. Both guests land on the same idea from different angles: good tools fail without good follow-through.WHAT YOU'LL LEARNNicol Russell points to a study out of NIEER at Rutgers that found teacher retention increased 23 percentage points when curriculum, assessment, and coaching functioned as one connected system rather than separate tools.Russell draws a line between teacher exhaustion and burnout, and she recommends that program leaders protect a set block of time each week where teachers have complete control over how it is used.Dr. Nneka McGee led a session at ISTELive that pairs Play-Doh with generative AI, using the activity to teach young learners about anthropomorphism while building creative thinking and questioning skills.McGee recommends introducing AI to young children through conversation rather than devices, citing both her own research on early learner readiness and guardrails like COPPA that already limit generative AI use for this age group.STORIES MENTIONED IN THIS EPISODEStory 1We Talk About Whole Children. What About Whole Educators? by Nicol RussellStory 2AI for Young Learners (Jump Start Guide) by Nneka McGeeSTAY CONNECTEDSubscribe to EdSurge newsletters at edsurge.com.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing:Nicol Russell, Chief Academic Officer, Teaching StrategiesDr. Nneka McGee, Founder, Muon Global

Today on This Week with EdSurge, we look at AI policy from two very different vantage points. One guest spent a year coding AI policy from more than 100 school districts across the country. The other helped write her own district's policy from the ground up. Listen as they both explore what happens when a policy leaves the page and is applied in an actual classroom.WHAT YOU'LL LEARNDistrict AI policies fall into five levels ranging from actively promoting AI to outright prohibition, and Jody Britten found that 44.3 percent of the 122 districts she studied land in the conditional, teacher-directed middle level.Only 3.3 percent of districts studied sit at Level 1, actively promoting AI with equity built into the design, while nearly a quarter restrict or ban AI outright.Most districts studied operate in states with official AI guidance, yet only 15.6 percent of those districts actually reference that guidance in their own policy.Tambra Clark found that writing her own district's AI governance policy in Birmingham required retraining her committee first, because knowledge in the room was too uneven to write a single line.STORIES MENTIONED IN THIS EPISODEStory 1Study: The National AI Policy Landscape in K-12 Education by Jody BrittenAI School Policy DatabaseStory 2Learn about the Voices of Change FellowshipSTAY CONNECTEDSubscribe to EdSurge newsletters at edsurge.com.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing:Jody Britten, Interim CEO, AI and Education StudioTambra Clark, Technology Integration Facilitator, Birmingham City Schools

This week, two educators arrive at the same question from very different directions. Nathan Kraai, Director of Innovation and Design Thinking at The Fenn School, explains what a trip to the record store taught him about artificial intelligence and learning. Court Shuller, a middle school English teacher and EdSurge Voices of Change fellow, joins from ISTELive in Orlando to talk about what their classroom has already decided about AI. Together, they ask what should still take real effort when a shortcut is only a click away.WHAT YOU'LL LEARNNathan Kraai connects the resurgence of vinyl records to what he calls a signal about how people are actually responding to AI, drawing on Jane McGonigal's trends versus signals framework from her book Imaginable.Nathan also explains why watching 9th graders wrestle with capstone speeches, despite having AI tools that could draft a polished version in seconds, left him hopeful rather than worried.Court Shuller, a National Board Certified educator and reading specialist, discusses how their district in New Jersey has responded to AI by returning to pencil and paper.Both guests weigh in on what students should still be required to do themselves.STORIES MENTIONED IN THIS EPISODEHow a Vinyl Record Resurgence Helped Me Understand the Future of AI in Education by Nathan KraaiWe Must Stop Using AI to ‘Level-Down’ Our Students by Court ShullerSTAY CONNECTEDSubscribe to EdSurge newsletters at edsurge.com.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing:Nathan Kraai, Director of Innovation and Design Thinking, The Fenn SchoolCourt Shuller, Middle School English Teacher and Reading Specialist

Every few months, another headline declares the college degree dead. This Week with EdSurge puts that claim to the test with two guests defending something slow against something fast. Rita Finkel of the Armory Foundation argues that a four-year degree still holds value at a moment when artificial intelligence can produce a credential adjacent skill in minutes. Then Cobretti Williams of the EdSurge Voices of Change Fellowship makes a similar case for writing, where an essay can take months to draft even though AI could produce one in eight seconds. Both guests land on the same idea: skip the time, and you skip the point.WHAT YOU'LL LEARNBachelor's degree holders earned a median of $81,800 in 2024, compared to $50,600 for high school graduates, a gap of 62%, according to the College Board's Education Pays 2026 report.From 2000 to 2025, workers with only a high school diploma consistently faced unemployment rates at least 2.3 percentage points higher than college graduates, according to a St. Louis Fed analysis.A McKinsey survey of more than 3,000 employers found that skills such as critical thinking and judgment remain in high demand from employers even as AI adoption grows.Cobretti Williams describes how the EdSurge Voices of Change Fellowship helps teachers develop their writing across multiple drafts, arguing that the growth between a fellow’s first and fourth essay only comes from doing the work yourself.STORIES MENTIONED IN THIS EPISODEWhy College Degrees Matter in the Age of AI by Rita FinkelEdSurge Voices of Change Writing Fellowship by EdSurge, featuring Cobretti WilliamsSTAY CONNECTEDSubscribe to EdSurge newsletters at edsurge.com.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing: Rita Finkel, Co-President, Armory FoundationCobretti Williams, Editor, EdSurge Voices of Change Fellowship

Microsoft just released the third edition of its global AI in Education report, and the numbers raise more questions than they answer. Ira Apfel sits down with Pat Yongpradit, General Manager of Global Education and Workforce Policy at Microsoft, on the floor at the ISTELive and ASCD Annual in Orlando. They dig into the delta between daily use and overall adoption of AI, a sharp drop in student optimism, and a persistent training divide between what leaders believe and what teachers and students actually experience. The conversation also turns to academic integrity and where the policy backlash against AI in schools may be headed next.WHAT YOU'LL LEARNDaily AI use lags far behind overall adoption, with just 33 percent of educators and 27 percent of students using it every day despite nine in ten having tried it.Student optimism about AI dropped 21 points in a single year, a shift Pat links to a predictable trough in the technology adoption curve.Three out of four students and half of educators report no formal AI training, even though seven in ten leaders believe their people are trained.Yongpradit explains why academic integrity worries students and educators for opposite reasons, and what that means for how schools should write AI guidance.STORIES MENTIONED IN THIS EPISODE2026 AI in Education Report, Published by MicrosoftSTAY CONNECTEDSubscribe to EdSurge newsletters.Find the latest education news at edsurge.com/news.Follow EdSurge on LinkedIn, X, Facebook, TikTok, and Instagram.HOST AND CONTRIBUTORSHost: Ira ApfelFeaturing: Pat Yongpradit, General Manager of Global Education and Workforce Policy, Microsoft