Title: How AI Transforms Learning - with Former Harvard Professor and Dean Stephen Kosslyn Resource URL: https://podcasts.apple.com/us/podcast/how-ai-transforms-learning-with-former-harvard/id1721313249?i=1000673255511 Publication Date: 2024-10-16 Format Type: Podcast Reading Time: 61 minutes Contributors: Stephen Kosslyn;Henrik Werdelin;Jeremy Utley; Source: Beyond the Prompt (Website) Keywords: [Artificial Intelligence, Learning Science, Active Learning Strategies, AI in Education, Cognitive Amplification] Job Profiles: Corporate Trainer;Learning and Development Specialist;Training and Development Manager;Chief Technology Officer (CTO);Chief Executive Officer (CEO); Synopsis: In this podcast, Beyond the Prompt hosts Jeremy Utley and Henrik Werdelin, and Active Learning Sciences Founder and Former Harvard Dean and Professor Stephen Kosslyn explores how AI is reshaping active learning as a cognitive amplifier to improve critical thinking, adaptability, and problem-solving. Takeaways: [Active learning is more effective than passive learning because it forces learners to use knowledge rather than just absorb it., AI serves as a "cognitive amplifier" by enhancing human strengths and compensating for memory limitations and weaker information processing., AI can support deep learning by providing varied examples, prompting critical thinking, and helping learners apply knowledge in different contexts., Custom AI-driven education can improve engagement by tailoring content delivery to individual needs and helping maintain motivation., Humanities are more important than ever in the AI era, as they help people develop contextual understanding, creativity, and adaptability.] Summary: Stephen Kosslyn, a leading cognitive psychologist and AI-driven education expert, discusses how AI can enhance learning rather than replace human educators. He defines active learning as “learning by using,” rather than simply absorbing information, which significantly improves retention and application of knowledge. Passive learning, such as lectures, often leads to 70–90% of content being forgotten within days. Kosslyn describes AI’s role as a cognitive amplifier rather than a co-pilot or collaborator. AI does not have goals or personal understanding but can boost human cognition by enhancing memory, critical thinking, and knowledge organization. He introduces the Cognitive Amplifier Loop (CAL), a structured way of using AI to refine thinking: Start with a goal, which can be vague at first. Use AI to refine or redefine the goal. Generate a response and analyze it critically. Adjust the prompt or goal based on AI feedback. Iterate the process to improve understanding and application. He highlights one of the biggest challenges in learning—the “transfer problem”, where students fail to apply knowledge learned in one context to another. AI can help solve this by providing diverse examples, prompting analogical thinking, and guiding learners through problem-solving frameworks. Kosslyn also shares his approach to personalized learning through AI. His company, Active Learning Sciences, develops educational programs that incorporate AI-driven teaser videos, knowledge assessments, personalized content delivery, and real-time feedback. The AI adapts lessons based on learners’ strengths and weaknesses, ensuring more effective and engaging instruction. For corporate settings, AI-powered training programs can enhance employee skill development, but Kosslyn warns against relying solely on extrinsic motivation (rewards and promotions). He emphasizes the importance of intrinsic motivation, which can be nurtured by making learning relevant, engaging, and socially connected. A key realization from his latest book is that the humanities are more valuable than ever. Literature, philosophy, and the arts help develop contextual reasoning, creativity, and perspective-taking, which are skills AI struggles to replicate. In a world increasingly influenced by AI, humanistic education remains essential for decision-making and problem-solving. Content: ## Introduction In a recent discussion, Stephen Kosslyn—former Harvard professor and dean—reflects on his transition from academic research in psychology and neurology to leading AI-driven educational ventures. Over decades, Kosslyn has championed a pedagogical approach he calls **active learning**, underscoring the distinction between mere exposure to information and **learning by using**. He argues that applying concepts in open-ended, real-world scenarios not only enhances retention but also fosters creative application. This conversation surveys Kosslyn’s career trajectory, his insights on AI as a **cognitive amplifier**, and the enduring role of the humanities in an AI-augmented educational landscape. ## From Academia to EdTech Leadership ### Early Academic Career After serving on the neurology faculty at Massachusetts General Hospital, Kosslyn returned to Stanford—his graduate alma mater—to direct a center for advanced behavioral science. When that effort proved less impactful than hoped, he joined a startup university initiative, helping to design curricula from the ground up. ### Founding AI-Driven Educational Startups Kosslyn then co-founded a program for working adults, focusing on skills less susceptible to automation. Today, he leads **Active Learning Sciences**, an organization that leverages AI to develop interactive, competency-driven educational programs worldwide. ## Defining Active Learning ### Learning by Using Kosslyn rejects the notion that active learning is simply “learning by doing.” Instead, he frames it as **learning by using**: learners pursue specific objectives by applying new material in debates, simulations, problem-solving sessions, or role-playing exercises. This engagement transforms information from passive content into a dynamic resource. ### Passive vs. Active Outcomes Research indicates that up to 70–80 percent of information encountered passively—through lectures or listening—can be forgotten within 24 hours. After three days, retention may fall to about 10 percent. In contrast, active use of material leads to deeper integration and greater readiness to apply knowledge creatively in unfamiliar contexts. ## AI as a Cognitive Amplifier ### The Cognitive Amplifier Loop Kosslyn introduces the **Cognitive Amplifier Loop**: a four-stage process—goal formulation, prompt construction, AI-generated output, and learner refinement. At each stage, learners exercise critical thinking (e.g., evaluating source credibility, assessing logical coherence) and creative thinking (e.g., divergent ideation, convergent selection). 1. **Goal Setting**: Define or refine the learning objective. 2. **Prompting**: Craft a prompt that directs the AI toward the goal. 3. **Review**: Critically evaluate the AI’s response. 4. **Refinement**: Adjust the goal or prompt, or iterate toward a polished result. ### Extending Working Memory By embedding detailed tables of critical-thinking categories or other reference material into the AI’s context window, learners effectively expand their working memory. Instead of struggling to recall dozens of evaluative criteria, the AI maintains them and applies them on demand. ## Challenges and Limitations ### Context-Specific, Open-Ended Problems Kosslyn emphasizes that humans currently excel at tackling open-ended tasks requiring nuanced context awareness—an area where AI still lags. While large language models interpolate well within their training data, they can struggle to **extrapolate** to entirely new contexts. ### The Transfer Problem Decades of learning-science research highlight **transfer**—the ability to apply knowledge from one context to another—as a persistent challenge. Classic experiments, such as the **radiation-treatment** problem and its military analogy, show that fewer than 30 percent of learners spontaneously transfer abstract solutions unless explicitly prompted. AI can mitigate this by presenting varied examples and guiding learners to recognize underlying principles. ## Early AI Experiences Kosslyn’s first forays into neural-network modeling date to the early 1990s. However, his appreciation for modern generative AI crystallized in late 2022, when he experimented with ChatGPT 3.5’s ability to pause for user input—enabling branching, interactive exercises essential to active learning. ## Implementing AI-Enhanced Instruction ### Four-Phase Learning Sequence At Active Learning Sciences, each class session follows a structured, AI-supported sequence: 1. **Teaser Video**: A short, counterintuitive clip—generally AI-generated—to spark curiosity. 2. **Pre-Assessment**: AI conducts an interview or quiz to identify what learners already know. 3. **Content Delivery**: Targeted instructional materials, tailored to address knowledge gaps. 4. **Active Learning Activity**: Interactive exercises graded and critiqued in real time by AI, with immediate tutorials for areas of difficulty. ### Motivation: Intrinsic and Extrinsic Drawing on **self-determination theory**, Kosslyn notes three intrinsic drivers: competence, autonomy, and relatedness. AI’s responsive personalization can calibrate challenge levels (competence), offer choices (autonomy), and support collaborative interaction (relatedness). In corporate environments, extrinsic incentives—such as clear career pathways—often complement intrinsic motivation. ## Fostering Deep Processing and Associations The most robust learning emerges from **deep processing**: paying focused attention, elaborating on material, and forging strong mental connections. AI can prompt learners to compare and contrast concepts, generate examples and exceptions, and practice deliberately on particularly challenging elements. ## The Imperative of Humanities Surprisingly, Kosslyn concludes that the humanities are more vital than ever. Engaging with literature, art, music, and philosophy cultivates the reflective and ethical capacities necessary to set meaningful life goals in an AI-saturated world. ## Conclusion Stephen Kosslyn’s vision of AI-transformed learning emphasizes active, context-rich engagement and the strategic use of AI as an extension of human cognition. By integrating critical-thinking frameworks, structured iteration loops, and motivational scaffolds, educators and learners can harness AI not as a replacement for human creativity but as a powerful amplifier of it. As we navigate this new educational frontier, the enduring insights of the humanities will remain indispensable guides.