# AI+Education Summit 2026: Scaling Human-Centered AI – What It Takes to Transform Learning for All Source: https://www.youtube.com/watch?v=JWKXCv1HXBQ Recap page: https://rapidrecap.app/video/JWKXCv1HXBQ Generated: 2026-02-19T18:13:06.706+00:00 --- ## Quick Overview The scaling of human-centered AI in education requires a multidisciplinary approach involving educators, social scientists, and technologists, focusing on measurable, human-centric outcomes rather than solely on technical capabilities or traditional metrics like test scores, to ensure equitable and beneficial transformation for all learners. **Key Points:** - The panel discussed the current moment as an amazing opportunity where AI in education is moving from possibility to reality, requiring a multidisciplinary effort. - Susan Athey noted that past technology adoption in education was often slow and that AI's current rapid adoption presents a risk of widening existing gaps if guardrails are not proactively implemented. - James Landay emphasized that human-centered AI design requires involving social scientists, humanists, and domain experts (like educators) from the start, not just engineers, to ensure societal impact. - Ian Chiu highlighted that early-stage educational AI companies like Newsela and Kiddom focus on augmenting teacher capabilities, like providing real-time data feedback, rather than replacing teachers. - Amanda Bickerstaff pointed out that most current AI usage is for mental well-being or belief offloading, not direct schoolwork, and stressed the need for clear guardrails to prevent negative outcomes. - The necessity of robust, multi-faceted measurement, including human-centric metrics beyond simple test scores, was identified as crucial for ensuring AI tools benefit all students equitably. - The panelists agreed that the focus must shift from simply building powerful AI to intentionally designing it to improve human skills and societal outcomes, requiring intentional design choices from the outset. ![Screenshot at 00:04: The panel title slide introduces the summit theme: "Scaling Human-Centered AI: What It Takes to Transform Learning for All," featuring the five main speakers.](https://ss.rapidrecap.app/screens/JWKXCv1HXBQ/00-00-04.jpg) **Context:** This video captures a panel discussion from the 2026 AI+Education Summit hosted by Stanford University's Human-Centered Artificial Intelligence (HAI) Institute and the Accelerator for Learning. The panel, moderated by Lewis Leiboh (Gates Foundation), featured experts including Susan Athey (Professor of Economics), James Landay (Professor of Computer Science), Amanda Bickerstaff (CEO, AI for Education), and Ian Chiu (Managing Partner, Owl Ventures). They discussed the challenges and requirements for successfully scaling human-centered Artificial Intelligence to transform learning equitably for all students and educators. ## Detailed Analysis The panel addressed the critical transition of AI in education from theoretical possibility to real-world application, emphasizing the need for a human-centered approach. Susan Athey warned that the rapid pace of AI adoption risks exacerbating existing educational inequities if proper guardrails and thoughtful design are neglected. James Landay stressed that successful human-centered AI requires interdisciplinary teams including social scientists and domain experts, not just computer scientists, to define what should be built based on broader societal impact, not just technical capability. Ian Chiu provided examples from venture capital, noting that successful educational technology often augments teachers by providing data insights, allowing teachers to focus on high-value human interactions. Amanda Bickerstaff highlighted the current reality where many young users employ AI for non-academic purposes like mental well-being and stressed the danger of choosing tools that lead to negative societal effects. A key theme was the difficulty in measuring meaningful educational impact, moving beyond simple test scores to assessing complex human skills and societal outcomes. The consensus was that intentional design, incorporating multidisciplinary expertise and robust evaluation, is necessary to ensure AI serves human goals rather than creating new bottlenecks or harms. ### Panel Introduction and Context - Lewis Leiboh introduces the panel (Susan Athey, Amanda Bickerstaff, Ian Chiu, James Landay) and the theme of scaling human-centered AI in education. ### Concerns about Rapid Adoption - Susan Athey notes the rapid adoption of AI and warns that without guardrails, it could widen existing gaps, referencing past technology rollouts that focused only on measurable metrics. ### The Need for Interdisciplinary Design - James Landay argues that designing for human-centered AI requires input from social scientists, humanists, and domain experts (like educators) to ensure the technology addresses real human needs and societal impact, moving beyond pure engineering metrics. ### Current Landscape and Teacher Augmentation - Ian Chiu discusses how current successful EdTech companies focus on augmenting teachers by providing data/insights to free up teacher time for high-value human interaction. ### Risks and Measurement - Amanda Bickerstaff highlights current uses (mental well-being, belief offloading) and the risk of locking in ineffective tools. She stresses the need for human-centric metrics and evaluation pipelines that look beyond simple test scores to assess societal impact and skill development. ### Call to Action - Susan Athey emphasizes that the current moment requires proactive, thoughtful decisions about measurement and guardrails, ensuring AI development serves human goals rather than creating new forms of harm or inefficiency. ![Screenshot at 00:02: Panelists and moderator introduced on screen with their titles and affiliations.](https://ss.rapidrecap.app/screens/JWKXCv1HXBQ/00-00-02.jpg) ![Screenshot at 00:15: Lewis Leiboh speaking about the current moment being a transition from possibility to reality in AI in education.](https://ss.rapidrecap.app/screens/JWKXCv1HXBQ/00-00-15.jpg) ![Screenshot at 00:35: Amanda Bickerstaff speaking on the importance of the current moment for democratizing the ability to create AI products.](https://ss.rapidrecap.app/screens/JWKXCv1HXBQ/00-00-35.jpg) ![Screenshot at 01:33: Lewis Leiboh discussing how the capability and speed of AI adoption differ from past technology waves.](https://ss.rapidrecap.app/screens/JWKXCv1HXBQ/00-01-33.jpg) ![Screenshot at 02:02: Susan Athey elaborating on the need for a 'gold standard' for evidence on human tutoring.](https://ss.rapidrecap.app/screens/JWKXCv1HXBQ/00-02-02.jpg)