What To Do When Research Lacks Answers

In March 2026, Stanford’s AI Hub for Education began compiling its “Evidence Base” on AI in K-12. The full collection of K-12 relevant papers now exceeds 1,600. As of November, Stanford researchers have identified 52 that feature systematic comparisons of outcomes in AI and non-AI conditions. Jeremy and Betsy talk with AI Hub for Education Director, Chris Agnew about the still emerging conclusions.Three key takeaways from the Stanford investigation:AI is better for students when they use it to support learning rather than simply producing results or finding answers;Even though AI can improve students’ perception of learning, it may not simultaneously improve their learning;How AI support is designed and delivered shapes its effects on learning. But it’s still unclear about which shape matters the most.Chris shares insights into how school leaders should think about these emerging observations. And Jeremy and Betsy have some pointed questions including: If what matters is how the AI tools are used, should we be pouring resources into AI for schools now?So you want to learn more?You can find all the details about Stanford’s research on AI in education at its AI Hub for Education. Here’s the most recent report, which we discuss in this episode.Also just released: the Hub’s research agenda; it’s a roadmap of questions the program aims to address.Stanford also share what’s known about AI-based tutoring here. And the team offers this advice for school leaders about questions they should ask when evaluating edtech tools here.Edtech Insiders did a nice job of summarizing the highs and lows of the Hub’s work on its substack here. They ask: What makes students want to use AI to learn rather than just get work done? No clear answers to that one yet.Check out the preprint of Patrick O’Neill’s analysis, “Presumed Effective: The Manufacturing of an Evidence Base for AI-in-Education Through Flawed Meta-Analysis” here.The Hechinger Report offers a very smart summary of O’Neill’s study here. And here’s a fascinating webinar on the study led by MIT’s Justin Reich .And for contrast: Here’s an approach that a Stanford researcher is taking to the question of whether AI is just helping students pass their medical licensing exams or helping them learn how to deal with (real and human) patients.Chris Agnew also had some delightful suggestions for us:For those in New York City, the one-man play, School Pictures, is playing through November at the Cherry Lane Theatre. It’s a one-hour musical that addresses parenting, inequality, and the anxieties of puberty in the New York City school system with humor and wit. “Not to be missed,” Chris says.And the beloved Kazuo Ishiguro novel, Klara and the Sun, is coming out as a movie starring Jenna Ortega and Amy Adams. Early reviews are mixed but its themes of loneliness in a world filled with artificial life resonate. Hosted on Acast. See acast.com/privacy for more information.

Why is deep change so hard in public schools? We know real learning doesn’t happen in tidy 50 minute blocks. But for the past 120 years, American schools have been captive to a “bell system” that is as relentless as an atomic clock. In this episode, Jeremy and Betsy explore the implications of the “Carnegie Unit” with researcher Dr. Brooke Stafford-Brizard of the Carnegie Foundation. Brooke shares how ongoing work might finally break the stranglehold that the Carnegie Unit has put on learning.Who is Brooke? Dr. Brooke Stafford-Brizard is senior vice president for Innovation & Impact at the Carnegie Foundation. She oversees Carnegie’s research and development activities and stewards cross-sector partnerships to accelerate the Foundation’s mission.Brooke has spent the past three decades bringing a holistic approach teaching to classrooms and schools, grounded in the science of learning and development. She got started as a middle school teacher in the Bronx. She co-founded an all-girls public charter school in Rochester, NY and later served as director of data strategy and evaluation at the New York City Department of Education’s District 79. She also served as vice president for research to practice at the Chan Zuckerberg Initiative. She earned her PhD in cognitive science in education from Columbia University, and is a Pahara-Aspen Education Fellow and a member of the Aspen Institute’s Global Leadership Network.Explore More The education research agenda of Carnegie Foundation for the Advancement of Teaching is ambitious and devoted to changing assessment from simply about grades to about demonstrating competency. Read all about it here.Brooke describes Carnegie’s work on “durable” skills. You can dive into the first Skills Progression: Collaboration, Communication and Critical thinking here. All nine are described here. Full reports will be out in 2028.Explore the history of the Carnegie Unit and proposals to change it by starting with this 2015 report, “The Carnegie Unit: A Century-Old Standard in a Changing Education Landscape.”XQ is a nonprofit organization affiliated with the Emerson Collective, founded in 2015 by Lauren Powell Jobs. Its Super Schools Initiative awarded $102 million to existing and new high schools in 16 states. These days, it says its work has reached 40,000 high school students in 150 sites. Check out this film, The First Class, that chronicles the creation of Crosstown High School in Memphis, Tenn.Earlier this year, the Stanford Social Innovation Review, published by Stanford University, hosted a series of 10 articles called “Renewing Public Education’s Purpose.” Check them out here.Consultancy Knowledge Works just released this report on developing a local policy framework for student centered learning. It also has this description of “Portrait of a Graduate.” More than half of the US states have developed some version of the portrait for shaping learning going forward. Hosted on Acast. See acast.com/privacy for more information.

Jeremy and Betsy discuss the recent podcast interview of Jensen Huang, CEO of Nvidia, by New York Times columnist and podcaster, Ezra Klein. Schools want to build "agency" in students. How will that happen? And what about the CEOs of the AI giants? How much agency do they really have? More to Explore: In this episode of Future Fluent, Jeremy and Betsy talk about a recent interview that New York Times columnist and podcaster, Erza Klein, did with Jensen Huang, CEO of Nvidia, which makes many of the chips used to power AI. Check out the full interview here. Huang was responding to comments by AI CEOs Sam Altman of Open AI, and Dario Amodei of Anthropic. Nerves were set on edge following reports that AI models had hacked into sites including one involved with Australia’s universal healthcare system. In this long essay on his blog, Amodei discussed the need to slow down AI. “I have worked on AI for the last twelve years because I believe it could dramatically raise the quality of human life,” he began. And then he added: “But over the last few months, I have become convinced that fully addressing the risks requires even more prudence — not just investing in risk prevention, but pacing the rate of capabilities advancement so that risk prevention has time to keep up.” Sam Altman went a step further in a post on X: “We welcome a federal framework that sets consistent safety requirements for frontier AI,” Altman wrote. “Pacing will be well worth this cost; no amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring.” You can hear more on Altman’s comments in this article on CNBC. Hosted on Acast. See acast.com/privacy for more information.

Can AI for teachers and students be built on a foundation of what we know about how students learn, namely on the learning sciences? And, if so, how can learning sciences infuse not just one AI product, but potentially all of them? Jeremy Roschelle and Betsy Corcoran kick off this third season of Future Fluent by talking with two people who are trying to change the “building blocks” of the AI programs used in schools. In this exclusive conversation with Priscilla Chan, cofounder and coCEO of the Chan Zuckerberg Initiative, and her colleague, Sandra Liu Huang, president of CZI’s Learning Commons division, Jeremy and Betsy explore what they are doing to retool the AI that is permeating edtech.Explore MoreSeptember marks the first anniversary of CZI’s Learning Commons, an AI “infrastructure” that any developer can use to build edtech tools. Companies (and yes, individuals, too), can sign up to explore the platform here: platform.learningcommons.org/login.The platform aims to have three components:* The Knowledge Graph, which aims to integrate materials such as rubrics, standards and even “misconceptions,” typical mistakes that learners make;* Evaluators, or ways to assess if AI material is aligned to learning science;* Agent Skills, or ways to help AI tools build learning science and strong pedagogy into whatever they’re building (say, lesson plans).Here’s a piece in EdSurge that describes the work. In it, Helen Crompton, executive director of the Research Institute of Digital Innovation in Learning at Old Dominion University, makes an important point: “Once a platform becomes part of the underlying infrastructure upon which many other products are built, it can begin to shape what information is available, what is considered authoritative, what gets measured and ultimately what kinds of educational experiences are created.”Here’s what Sandra and Priscilla are reading:* “Build: Investing in America’s Infrastructure“, by Sadek Wahba, managing partner at private equity firm, I Squared Capital.* School Rethink 2.0, edited by Frederick M. Hess, Michael B. Horn, and Juliet Squire* The Correspondent by Virginia Evans, which, as Priscilla says, brings home the idea that parenting takes an entire community. Hosted on Acast. See acast.com/privacy for more information.

What do wizards have to do with Doritos or trout? In this wrap up episode to Season 2, Jeremy & Betsy talk about the metaphors and insights their guests shared this year to make sense of AI. Guests included researchers, writers, educators, builders & influencers:  Ethan Mollick, Pat Yongpradit, Inge Molenaar, Erin Mote, Justin Reich, Melina Uncapher, Elvira Salazar, Margaret Honey, Mike Taubman and Babak Mostaghimi. And the theme running through it all: Sure, the technology is changing rapidly. But even more important: Our sense of when to use AI– and when not to use it–is morphing, too. Finally, Betsy & Jeremy hint at what they will explore when they kick off Season 3 this autumn. Got thoughts? Suggestions? Questions? Drop a note on LinkedIn. Hosted on Acast. See acast.com/privacy for more information.

The backlash against screens (and technology in school broadly) is growing, and with good reason. In this episode of Future Fluent, Betsy and Jeremy explore with Dr. Babak Mostaghimi of Learner Studio what it takes to build technology, including AI-based tutors, that bring people together rather than isolating them. The bottom-line question: Learning should help people thrive: How could AI be designed to do that?Oh, it’s hard to be human these days! Start with reading all about it. Babak Mostaghimi is part of the Learner Studio team. The team published a collection of pieces (including one by Betsy!) called “Bringing the Future Into Focus.” And check out Mostaghimi’s piece, “How can AI be used as tool for learning, beyond more efficiency toward current outcomes?”   Michelle Culver created The Rithm Project, a nonprofit, to “think out loud about what it means to be human–together–in an AI age.” Her substack is here. Julia Freeland Fisher, director of education research at the Christensen Institute, has written extensively on the value of networks. Check out this report: Students’ hidden networks: Relationship mapping as a strategy to build asset-based pathways.We talked about the work going on at OKO Labs in teaching math. And the work reimagining robots that’s going on at IDEO. McKinsey Global has been reporting on the future of work here:“Jobs lost, jobs gained: What the future of work will mean for jobs, skills, and wages.” The World Economic Forum has this set of stories about the role Artificial Intelligence may play in the world here. Dan Meyer, who never pulls his punches, put up this challenge in his Substack: “I’d like to see from everyone engaged in this discourse: one (1) stationary video of a full classroom session.” The challenge is on!  Hosted on Acast. See acast.com/privacy for more information.

Do we – or don’t we? Teachers are both under pressure to press 'pause' on AI in the classroom and to prepare students for the future with AI. On this episode of Future Fluent, Jeremy and Betsy connect with Mike Taubman, a veteran teacher at Uncommon School in New Jersey to explore how he's bringing decades of hard-earned pedagogical insights to bear in guiding his students into AI usage. As a long-time English teacher and career advisor, nothing is more precious to Taubman than face time with students. But he also wants them to be able to go under the hood and understand the assumptions and mechanics of AI. His solution: To start building an “AI Driver’s License” for high school students. Taubman shares what's working--and what's under construction--along with why Newark's Mayor Ras J. Baraka recently stopped by Taubman's class. Spoiler alert: Taubman wasn't the one who invited him.  LEARN MORE!Mike Taubman got his start as an English teacher–so, no surprise, you can follow his work by reading both what he writes and what he likes to read. Taubman writes frequent updates about his classes and other experiences on his Substack, AI Waypoints. (Browse here for details about the “AI Driver’s License” curriculum Mike and fellow teacher, Scott Kern are developing as well as Mike’s exchange with protesters at the 2026 ASU-GSV meeting.) The NYTimes wrote about the AI Driver’s License here.  Check out this fabulous story about how two of Mike’s students convinced the Mayor of Newark, NJ, Ras J. Baraka, to visit their school and managed the whole visit. Here’s Mike’s description of the visit, too. Books recommended by Taubman include: The Lifecycle of Software Objects by Ted Chiang, a Hugo Award winning science fiction novella from 2010 that describes the “raising” of digital entities. It’s included in a compendium of Chiang’s work, Exhalation (2019). Ministry for the Future by Kim Stanley Robinson, published in 2021, is a fav of both the NYT’s Ezra Klein and Barack Obama.  In Search of Deeper Learning: The Quest to Remake the American High School, by Jal Mehta and Sarah Fine from 2019, is an overview of the state of US high schools. And of course, the 2013 movie, “Her,” by Spike Jonze. We also talked about this (meta) study from Stanford University on 800 academic papers on the relevance of AI for K-12 education. Finally, you can also watch this segment on NBC’s Future Education about Taubman’s colleague, teacher Scott Kern, who built chatbots to support his students.  Hosted on Acast. See acast.com/privacy for more information.

What’s been the most powerful learning experience that you've had? Dr. Margaret Honey has helped build remarkable learning experiences–starting with the television show, The Voyage of the Mimi, through her work at the New York Hall of Sciences and most recently with the Scratch Foundation. Through it all, she’s held fast to several principles, starting with: Never fake it. And center activity around children’s curiosity not around rubrics or assessments. In this episode, Margaret shares with Jeremy and Betsy the triumphs, challenges and hard-won lessons learned of building memorable experiential learning environments–along with what changes in an AI-saturated world. (And, we also learn why actor Ben Affleck knows so much about humpback whales!) LEARN MORE!The Voyage of the Mimi (with Ben Affleck) was a 13-episode television program created in the mid 1980s. (Here’s episode 1.) A crew of the ship, Mimi, explored the ocean, to carry out a census of humpback whales. In The Second Voyage of the Mimi, archaeologists searched for a lost Mayan city. Here’s a video short on the Connected Worlds exhibit at the New York Hall of Science. (Better: Check out the exhibit at the museum!) Scratch, a free, nonprofit coding community and environment for children, is supported by the Scratch Foundation. (Start here if you’re considering a family membership; here if you’re an educator.) Xperiential, a collaboration between Pixar and Khan Academy, is a project-based learning approach aimed at inspiring students to explore careers through storytelling and design. Jeanne Bamberger’s 1995 book, The Mind behind the Musical Ear, explores how children develop “musical intelligence.” Creative Acts for Curious People: How to Think, Create, and Lead in Unconventional Ways, by Sarah Stein Greenberg (2021) includes both stories and innovative exercises to build creative leadership. Hosted on Acast. See acast.com/privacy for more information.

History is chock full of new technologies developed with good intentions. But if we’ve learned anything over the past few decades, it’s that doing research and designing products has layers of complexity. It isn’t enough to just build tech for others – we have to build it in close partnership and community with those who will use it. In this episode of Future Fluent, Betsy and Jeremy talk with Dr. Elvira Salazar, a life-long educator, passionate devotee of STEM education and NASA, and now the Director of Online Learning & Technology for Latinos for Education. They’ll talk about what the AI community gets right – and gets wrong – in the rush to build the next great thing. Learn more!To explore more of Dr. Salazar’s work, a great place to start is the Latinos for Education website. She also contributed this piece, “Learnings from the Front Lines on Redefining Leadership for the Age of AI,” to EdSurge. Dr. Salazar described the work of CLEAR, or the Center for Leadership Equity and Research. You can explore the groups work as well as its AI initiative at Clearvoz.com This story, “AI Leaves Some Students Lost in Translation,” explores in more detail some of the promise and challenges of AI development for the Latino community. (You can also try out the Playlab app developed by the group, “Elevating your Speaking,” a tool that parents can use to support their students’ language development skills, here.) Stanford University professor, Dr. Sanmi Koyejo, discusses his white paper about how AI is leaving non-English speakers behind here. The full report from Dr. Koyejo and his team is here: “Mind the (Language) Gap: Mapping the Challenges of LLM Development in Low-Resource Language Contexts.”  Hosted on Acast. See acast.com/privacy for more information.

Why are the “myths” about how we learn sometimes more powerful than what research has delivered? In this episode of Future Fluent, Jeremy and Betsy interview Dr. Melina Uncapher, a neuroscience researcher and learning science practitioner who has spent her professional life trying to help educators and researchers work together. They plunge into one of the great myths of learning–that we have different “styles” of learning–and then ask the question: Can AI make it possible to weave learning science into the classroom? Betsy puts a challenge to Melina.Learn more! Here’s where you can explore some of the ideas around learning science and AI that we discussed with Dr. Melina Uncapher. “Strategic Surprise and the Future of Educational R&D” is an openly available white paper coauthored by Drs. Melina Uncapher and Jeremy Roschelle. The core point: Generative AI is catalyzing rapid shifts in learning practices that far outpace how we have traditionally responded to changes. This paper proposes to re-architect education R&D to be more agile, responsive, and, hopefully to create breakthrough potential. Designed for Brilliance is a workbook resource for teachers on how to apply learning science by Dr. Uncapher. At the moment, it’s available as a downloadable PDF. Check out Chapter 2: The top 7 Neuromyths, Busted.  On the broad topic of using evidence-based practices in the classroom, take a look at the Accelerate, Transform Scale Hub launched recently by Digital Promise and SRI. Background details here. Why good ideas stall in education – an opinion piece by Dr. Uncapher and Nat Kendall-Taylor in the 74Million. AI Snake Oil: What Artificial Intelligence Can Do, What It Can’t and How to Tell the Difference, by computer scientists Arvind Narayanan and Sayash Kapoor, from Princeton University Press, is well-worth a read. And if it all seems like a bit too much, check out the Pixar movie, Inside Out, which, as Melina says, puts the principles of learning science into animation!  Hosted on Acast. See acast.com/privacy for more information.