When Code Is No Longer the Challenge

Aligning learning, assessment, and teaching with the realities of AI-assisted programming

Reimagining CS Education

Agentic Engineering, Authentic Assessment, and Human-Centered Design for Learner Agency

This year, my invited talk series will argue a framework for how we might reimagine computing education around agentic engineering, authentic assessment, and human-centered design. Contact me at kevinl@cs.uw.edu to arrange a talk.

Computers have always been good at giving answers; generative AI makes answers nearly frictionless. For computing education, which has long treated producing code as evidence of learning, this raises a question about what we value and whose judgment we are cultivating. In software engineering, the bottleneck is shifting from writing code toward architecting, specifying, verifying, and taking responsibility for code no one wrote by hand. Yet the curricula, assessments, and professional norms emerging around this shift are being designed largely in and for a narrow slice of the world’s learners and workplaces. Drawing on a synthesis from the Consortium for Generative AI in CS Education and on classroom experiments at the University of Washington, this talk offers five paired frames—questions and answers, past and present, depth and breadth, learner and agent, self and society—for reimagining computing education. Along the way, it considers how students can learn to interrogate the design assumptions built into everyday technologies, how assessment can measure the learner rather than the agent, how “everyone can code” changes who computing education is for, and how community-engaged projects and dialogue can rebuild the agency that frictionless tools can erode.

If you would prefer to present, adapt, or extend these ideas yourself, the slides are licensed CC BY-NC-SA 4.0.