The professors and long-time UW collaborators worked on the new course over eight months and developed lessons and resources from a cautious viewpoint to better understand the risks and rewards of a rapidly evolving technology.
By Michael Grass
Center for an Informed Public
University of Washington
University of Washington Center for an Informed Public faculty members Jevin West and Carl Bergstrom have introduced a new course that they think every first-year college student should take. Modern-Day Oracles or Bullshit Machines?, a humanities course about how to learn, work and thrive in a world increasingly dominated by artificial intelligence, was designed for students with and without a technical background.
“Our instructor guide provides a choice of activities for each lesson that will easily fill an hour-long class,” Bergstrom, a UW Department of Biology professor, observed in an introductory Bluesky thread co-authored with West, an Information School professor and CIP co-founder.
The entire course is available freely online, and built around 18 online lessons, each which take 5-10 minutes. Each lesson “illuminates one core principle,” including “Autocomplete in Overdrive” (Lesson 1), “From Voice Cloning to Shrimp Jesus” (Lesson 7) and “Your Own Private Truman Show” (Lesson 17), hitting on themes of misinformation, creativity, and authenticity.
Bergstrom and West have published accompanying videos for the first two course lessons and additional videos will be made available on a rolling basis on the course website throughout the next few months.
The professors and long-time UW collaborators worked on the new course over eight months and developed lessons and resources from a cautious viewpoint to better understand the risks and rewards of a rapidly evolving technology.
From left to right, UW professors Carl Bergstrom and Jevin West.
“Large language models are both powerful tools, and mindless — even dangerous — bullshit machines. We want students to explore how to resolve this dialectic,” Bergstrom and West wrote. “We marvel at what LLMs can do and how amazing they can seem at times — but we also recognize the huge potential for abuse, we chafe at the excessive hype around their capabilities, and we worry about how they will change society.”
Bergstrom and West noted that they didn’t want to lecture “at” students about the risks of generative AI and the large language models they’re built on. “We don’t think that works,” they wrote. “Students use LLMs already; we want them to discover the problems for themselves.”
Modern-Day Oracles or Bullshit Machines? is a follow-up to Calling Bullshit: Data Reasoning in a Digital World, a course Bergstrom and West first co-developed in 2017 at the University of Washington. The data reasoning course became one of the most popular at UW and has served as a model for similar classes at hundreds of colleges and universities across the United States and around the world. The course also led to a book, “Calling Bullshit: The Art of Skepticism in a Data Driven World,” co-authored by Bergstrom and West and published by Penguin Random House in 2020.
Bergstrom and West hope the new course will be just as far reaching but note that the lessons and resources are still a work in progress, so they’ll be “adjusting and adapting and revising in the months to come” based on user feedback.
Learn more about Modern-Day Oracles or Bullshit Machines?
Michael Grass is the Center for an Informed Public’s assistant director for communications.