Reflections on Future Learning in Education

How do we reimagine college education, and what role will generative AI/LLMs play? Dr. Megan Ma and I taught 30 students from Stanford University School of Engineering, Stanford Law School and Stanford University Graduate School of Business in our “Building for Impact: Generative AI and Law” accelerated class this quarter. We lectured on foundational concepts and tools in generative AI/LLMs, and on a product management framework to build startups. Simultaneously, six students teams – each with CS, law and business students – developed generative AI startup ideas in the legal space. They then competed in an open Stanford law x LLM hackathon we organized. Our student teams built generative AI solutions as follows: (i) AI-powered patent due diligence (ii) Help veterans navigate the medical claims process (iii) Due diligence tool to vet legal docs and economic expectations (iv) Lawyer negotiation simulator to train junior lawyers (v) Automated due diligence report from M&A diligence materials, and (vi) AI-powered product compliance tool for healthcare products. A bonus: out of 70 teams in the hackathon, 3 of our student teams were top 10 finalists and 2 of them won the top overall prizes. This was my 8th year teaching this class at Stanford University, in which we’ve experimented widely on teaching methods and teaching AI. My thoughts on reimagining the future of college education: 1. Generative AI accelerates the learning process Both technical and non-technical students could immediately use generative AI to develop startup ideas into working prototypes. Many students had limited prior knowledge of gen AI. This democratizes the process of turning class concepts into real world solutions, and at a velocity I have never seen before in my classes. Other educational fields should also creatively embrace gen AI to accelerate learning. 2. Experiential learning—applying course content to real world solutions—accelerates mastery of concepts Lectures and other unidirectional teaching methods (such as Socratic method in law schools) are passive methods of yesteryear. With experiential learning, students understand and engage with course content more deeply. Students must master new concepts, build using these concepts, and evangelize them. In class, students showed urgency in mastering new concepts as they incorporated them in building their generative AI startups. This methodology can be used across a wide array of disciplines. 3. Break down departmental silos to unleash student creativity The exchange of disparate ideas between students of different educational backgrounds leads to unforeseen learnings. In our classes, we’ve consistently seen interdisciplinary teams create unexpected new concepts and startups. Our students always value interdisciplinary learnings. Megan and I thank our incredible students, class sponsor CodeX, The Stanford Center for Legal Informatics, its Executive Director Roland Vogl, and our guest speakers/coaches.

Yesterday, I developed an AI tool that crafts annotated bibliographies. As an educator, why would I do this? Here's why: 🤖 AI Literacy: Students are already using AI to assist with their writing tasks. By formally incorporating AI into the course, my goal is to educate them about the capabilities and limitations of LLMs – like AI 'hallucinations.' 🎯 Critical Engagement: My assessment criteria have evolved. I'm no longer assessing their ability to gather and correctly summarize information. Instead, I encourage students to critically evaluate the AI chatbot's output and interrogate their sources, analyzing how each article contributes to their research questions and goals. 🔀 Process over Product: The journey of learning is now in the spotlight. It’s about understanding steps students take in their research exploration, encouraging them to value the research and critical thinking process as much as the final product. 💻 Technology Proficiency: Proficiency with AI tools is becoming as fundamental as traditional research skills. By integrating AI into assignments, students gain hands-on experience with LLMs. To be clear, the annotated bibliography remains a staple in my course. However, I've shifted our focus from mere information gathering to in-depth critical analysis. Students now perform comparative analyses, contrasting their own evaluations with AI-generated content. In addition, each assignment now includes a self-reflection component. Students ponder over their process of selecting articles, how well each source fits into their research framework, and what they have learned from their interactions with AI. 🔗 Explore the annotated bibliography generator and share your thoughts: https://lnkd.in/gHUCKsCG (P.S. Be patient when using this tool - it needs time to "think") . . . #educationinnovation #ailiteracy #futureoflearning #chatgpt #aiforeducation #llms

Hot off the press: I'm proud to share Learn Capital’s 2025 Outcomes Report—an effort I helped lead, and one that reflects our evolving view on how AI can reshape learning for the better. Read the report: https://lnkd.in/gMstPX_D We need to optimize for the human experience, not just efficiency! Lately, I’ve been hyper-focused on what it means to build systems that truly serve all learners. We know AI holds incredible promise—to accelerate mastery, personalize learning, and connect people to real opportunity. But it also presents challenges we can’t ignore. At Learn Capital, we believe this is one of the defining issues of our time. And we’re working to ensure that the future of AI in education expands access, builds agency, and improves outcomes at scale. Our strategy focuses on three core frontiers: 1. Skill Delivery – foundational literacy, numeracy, creativity, and wellness 2. Labor Market Optimization – converting skills into income and mobility 3. Learning Infrastructure – AI-powered systems for lifelong learning We’re seeing powerful signals across our portfolio: - Amplify: +20 point reading gains in underserved districts - Brainly: 85% of AI tutor users improved grades - Ascent: $1.6B in income gains through outcome-based financing - Brilliant.org: 4× improvement in problem-solving - Polygence: 6× higher acceptance to top-tier universities - Andela: 77% income growth for engineers upskilled through AI-driven pathways Grateful to be spending my time and energy working to leverage AI as a force for improving lives—and building a future where learning is more human, more equitable, and more effective. Also, a big shout-out to my colleagues, Vinit Sukhija, Rob Hutter, Greg Mauro, and Sierra Espinosa. Also, huge thanks to our strategic communications partner, Matt W Gore, Nathan Wallace, and the OPTIO team for making this sing. Would love your thoughts. #AI #EdTech #ImpactInvesting #FutureOfWork #LearningEquity #HumanDevelopment #LearnCapital

This student is using OpenAI's GPT-4o to get real-time personalized audio explanations. He's speaking into an iPad to have this advanced AI tutor instantly break down complex topics. What AI can do is quite amazing. Do you think this is a glimpse into EdTech's future potential? Lecturing could be done differently, allowing educators to focus on facilitating rich discussions and applied learning. At the same time, AI reinforces core concepts through interactive practice. ➔ Or students mastering subjects at their own pace by simply asking GPT-4 for guided walkthroughs. ➔ The ability to personalize learning and provide on-demand tutoring at a massive scale is staring us in the face. But like any transformative technology, AI in education has some serious pros and cons. On the plus side, it could considerably enhance learning accessibility and quality. Supplementing human instruction with 24/7 virtual tutors could help bridge gaps where expertise or resources are limited. And honestly, routine knowledge transfer may be better suited for AI to free up educators to cultivate higher-order skills. The downsides, however, can't be ignored. There are huge hurdles around cost and equitable access to the required tech and internet infrastructure. A lack of human input could also hamper critical thinking development. We must also be vigilant about AI perpetuating societal biases through the data it's trained on. Ultimately, I believe EdTech like GPT-4 is too powerful an opportunity to ignore—but it would have to be implemented smartly with plenty of guard rails. We have a real opportunity to create a future of learning that allows greater access, more meaningful interactions between faculty and students, and a more efficient delivery model. However, we need to thoughtfully augment human expertise and pedagogy, not hand over learning to AI.

The boundaries of traditional education are being redefined! 🎓 As universities and colleges increasingly adopt virtual reality (VR), Meta is stepping up, providing VR equipment and resources to 15 U.S. universities. Their aim? To make classroom education more immersive and interactive. 🔍 What I observed: ◽️Hands-on learning. Experts believe VR and the metaverse could revolutionize hands-on learning, though there are challenges to overcome. ◽️Diverse applications. From teaching "soft skills" at Stanford to language immersion at Arizona State University, the potential applications are vast. ◽️Equity in education. VR can offer underprivileged students experiences previously out of reach - imagine a virtual trip to the Louvre or NASA! ◽️Research & development. While the adoption of VR in classrooms is still in its infancy, research supporting its efficacy is growing. Meta, for instance, is investing in understanding the benefits and risks of these technologies. However, it's not just about technology. It's about equipping educators with the right tools and ensuring students remain engaged. As Nick Clegg from Meta puts it, VR can't replace teachers but can provide them with innovative tools to enhance the learning experience. As we look ahead, the convergence of VR and AI promises a personalized education experience, tailored to each student's needs. The classroom of the future is not just a physical space but a blend of reality and virtual realms. Are you excited about this educational evolution? Let's embrace the future of learning! - #Education #VirtualReality #Metaverse #Innovation

Exciting news! 🚀 This week, Washington became the fifth state to unveil a strategy for integrating AI into public schools. 🙌 Here's why I think this sets a great precedent for other states/countries: 🤖 Human → AI → Human: Students and educators are encouraged to always begin with human inquiry, apply AI tools, and then evaluate and edit results. Focusing the narrative on human-centered approach ensures we maintain the integrity of the learning experience. 🚀 Empowering Future Innovators: AI is here to stay. This initiative encourages the embedding (vs. banning) of ethical AI use in K-12 education. This ensures students are not just passive users but informed creators and critics of AI technologies, ready to navigate a future where AI is likely much more ubiquitous. 🛡️ Data Protection & Privacy: Prioritizing the safety and privacy of student data is a big piece of AI adoption and literacy. We will never be able to establish trust among educators and parents in AI unless we ensure that student information is safeguarded. Arguably the EU is ahead of the US in this regard but it’s great to see data privacy is becoming an important part of the conversation in the US as well. 🌍 Equity and Inclusion in AI: There already are huge gaps in accessibility to meaningful AI education and so I appreciate multiple callouts in the document to ensure AI education is accessible to every student, breaking down barriers and promoting inclusivity. 🚀 Professional Development for Educators: The initiative recognizes the importance of empowering teachers with the knowledge and tools to effectively integrate AI into their teaching practices. 🤔 Critical Thinking and AI Ethics: Students are encouraged to engage with AI critically, understanding the algorithms and data that power these technologies. There are real risks and biases that come with GenAI and so it’s great to see that students are encouraged to question (and shape) the impact of AI on society. 📈 Real-World Applications: It's important to remind students that despite risks there are real positive use-cases of these technologies. And hopefully getting them excited to use GenAI meaningfully in their careers. 💡 Creative Problem Solving: By understanding AI's capabilities, students are equipped to leverage technology for creative innovation. This focus on creativity ensures that the next generation is ready to use AI in novel ways (that the non-AI natives might not think of). 👨🏫 Community Engagement and Support: Parents, families, and the wider community need to be involved in understanding AI's role in education. This inclusive approach ensures a collective effort in navigating the AI landscape. AI literacy is what we stand for at ai LaMo so huge kudos to Office of Superintendent of Public Instruction and Chris Reykdal for championing such a visionary approach! #EdTech #AIineducation #WashingtonState #GenAI #education #edtechstartup

Oddly enough, in pondering #AI in #education lately, I’ve been thinking about John Dewey, who wrote: “the most important factor in the training of good mental habits consists in acquiring the attitude of suspended conclusion, and in mastering the various methods of searching for new materials to corroborate or to refute the first suggestions that occur. To maintain the state of doubt and to carry on systematic and protracted inquiry ― these are the essentials of thinking” Suspended conclusion! Generative AI couldn’t be more opposite. Conclusions come in nanoseconds, hot out of the gate. Maintain a state of doubt? AI produces outputs with air of certainty (even when the factual basis is dubious). The later writings of John Dewey explore the nature of inquiry, which to Dewey, is the most essential form of “critical thinking” to be learned. Inquiry is all about opening up the space between means and ends. Slowing down. Interrogating presumptions. Searching for unforeseen connections. GenerativeAI may indeed help teachers or students produce this lesson plan or that assignment quickly or artfully. But who will help students to suspend conclusion, to not only corroborate -- but to refute initial thoughts which come immediately in near-final form? Thus I propose as a driving question for the design of learning activities that use AI: As the “doing” in “learning by doing” is automated, how do we hold space for inquiry?

Friday was an incredibly rewarding professional learning day with our high school staff! Our discussion on how teachers and students can use AI as a co-pilot, not a crutch, was thoughtful and eye-opening. We started with a Curipod Vocabulary Hustle review and dove into prompt engineering using the PREPARE framework from Dan Fitzpatrick. Teachers developed their own creative prompts and tried them using multiple LLMs. Next, we explored tools like fobizz which even the most cautious and skeptical teachers felt comfortable using due to its data protection measures. To shift to a discussion on student use of AI, we reviewed a modified AI Assessment Scale initially developed by Leon Furze and the AI for Education curriculum. Teachers are eager to create a plan to introduce students to AI and have honest conversations about how it can be a helpful and harmful tool based on its use. Finally, we ended with a low-tech “Gots & Wants” reflection activity. Teachers got a lot out of this session, and I did, too! Next, we will provide more time to explore tools and plan how to move forward. What I learned from this experience: 1. We cannot expect teachers to do this work independently. Administrators need to provide high-quality training and time to explore these tools. 2. Having an honest conversation about the possibilities of AI (both exciting and terrifying) goes a long way in building trust and understanding. Although AI excites me, to prepare for this, I listened to podcasts and read critical articles on AI to present a realistic approach. 3. Don’t focus on fancy AI tools. Center discussions around good instruction and the ethics and implications of AI in our schools. Thank you to the incredible teachers who engaged in this session. I’m looking forward to our next steps!

Reading recent posts about and reactions to Google’s ISTE announcement brings back memories from my classroom. This is what I imagine - I give my students a group project. They’re collaborating, discussing, problem-solving. BUT, I’m sitting at my desk, completely blind to their conversations. That’s essentially what Google just offered educators with Gemini. Larisa Black, Arman Jaffer, and Matt Miller are spot-on about the missing piece: teacher visibility into student AI interactions. As someone who spent 18 years as an educator, I know that often the most learning happens in the messy middle—in the questions they’re afraid to ask out loud, the misconceptions they’re working through, and the “aha!” moments when concepts finally click. When students chat with AI, those insights are LITERAL gold for teachers, but only if we can see them. Here’s what I wish I could tell every educator and district leader still pondering Google’s announcement: The shiny new tools aren’t the question. The question is: Will this help teachers teach better and students learn better? 👀 Can teachers see when a student is struggling with a concept for the third time? ❗️Are teachers alerted when a conversation takes a concerning turn? 🔎 Can teachers identify the student who’s actually engaging deeply vs. the one just getting quick answers? My students needed me to understand their thinking process, not just their final products and answers (truly - process > product). The same is true today with AI. I’m not anti-Google—their tools have transformed how we work (and I love them). But there’s a fundamental difference between AI built for productivity and AI built for pedagogy. Consumer AI might ask: “How can I help you get this done faster?” Educational AI asks: “How can I help you learn this better?” That distinction shapes everything—from safety features to teacher controls to how learning insights are surfaced. What are your immediate thoughts? #aiineducation #aiforeducation 📸 my students in an active learning progress!

The biggest barrier in education isn't kids 'cheating' with AI - it's helping teachers learn to teach with it. 🍎 This summer, I've been working alongside students and teachers, listening and learning. Many conversations focus on "preventing cheating," but I think we're approaching this wrong. Students are naturally curious and resourceful - that's a strength we can build on. The real opportunity lies with educators: helping them feel confident and supported as they explore new ways to engage students through AI integration. This transition is crucial because it removes the stigma around AI use and transforms it into educational opportunity to work with AI as a thought partner - a skill that will be essential for graduating students. 🎯 That's why Eric Howe and I created "The AI Lesson Architect," a Custom GPT designed to help teachers create lesson plans that thoughtfully integrate AI. Here's an example that excited us: We developed a lesson on Immigration, Industrialization, and Social Movements (1870–1900). Instead of a traditional approach, students step into the shoes of people their own age from that era - new immigrants, young factory workers, teen labor activists. They create diaries from the perspective of factory workers, developing empathy and making personal connections to history. (this is one of a million examples that can come alive with a thoughtful creative approach to working with AI as new tool) When teachers feel supported and equipped, incredible learning happens. The future of education is about empowering educators to create experiences that inspire deeper understanding and connection. 🏫 I'm curious: What support would be most helpful as you explore AI in your teaching or learning environment? If this resonates with you, I'd love your help sharing this with the educators, parents, and administrators in your network. You can find The AI Lesson Architect GPT link in the comments below - feel free to try it out and pass it along. #EdTech #TeachingWithAI #Education #Teachers #AI #LessonPlanning #StudentEngagement