Overview
Watch a Stanford University seminar exploring how language models can serve as temporary learning aids while avoiding over-dependence. Associate Professor Tim Althoff from the University of Washington discusses strategies for using AI collaboration to develop social and communication skills through structured assistance, practice, and feedback. Learn about research findings from multiple randomized trials involving over 150,000 participants that demonstrate how language models can effectively support skills like cognitive reframing, emotional regulation, and conflict resolution while encouraging eventual independence from AI tools. Discover insights from Althoff's work in AI and mental health services that has impacted millions and influenced federal policy. The talk emphasizes balancing the benefits of AI assistance with the importance of developing genuine human capabilities and critical thinking skills.
Syllabus
Stanford Seminar - Language models as temporary training wheels to facilitate learning
Taught by
Stanford Online