Course description

Over the past half-century, computation has woven through our material culture, our society, and our way of thinking. Much more recently, neurally inspired machine learning has begun to enable computers to perform cognitive tasks that most people have thought of as exclusive to humans: understanding, judgment, and perhaps even creativity. These developments can be understood on many timescales: from a zoomed-out, ecological perspective spanning billions of years, from a cultural perspective as the most recent development of a scientific and technological program articulated by Francis Bacon in the 17th century, and as a reunification of neuroscience and computer science, which share common roots from the 19th through the mid-20th century.

Companies, governments, and institutions are rushing to explore and exploit the possibilities that AI opens up. The range of potential outcomes is vast, from “Black Mirror” dystopias– or outright extinction– to utopias of ecological recovery, equity, and abundance. Those predicting one or another end of this range tend to either be advancing ideological agendas, or claiming that large-scale trends or “laws” will inevitably produce certain outcomes– though often agendas are offered in the guise of predictions. While we will examine large-scale trends and patterns, the premise of this course is that the future is not settled, and that our own design choices will matter.

As the engineers, policy-makers, economists, politicians, scientists, and humanists of the century ahead, navigating this landscape will require that we transcend traditional fields of study to try to make better sense of where we are, how we got here, and how we might move forward. While there are no formal prerequisites, students taking this course should be prepared to read and discuss primary sources and difficult writers from widely varying intellectual traditions, including philosophy, neuroscience, art, critical theory, economics, the history of ideas, and science and technology studies. Students will also benefit from having a practical grounding in computer science and data analysis; if this background is lacking, they must at a minimum be unafraid of plots.

The territory we will cover is very broad, but this is not a survey course, as it delves into selected topics in depth, and the choice of material and sources necessarily reflects a point of view. While modern China and Russia are touched on with regard to AI and the emerging sociotechnical landscape, most of our historical focus will be on the West.

Regardless of personal beliefs, students must be comfortable talking about race, gender and sexuality, religion, and other sensitive topics in a way that is inclusive, curious, and undogmatic.


Ten 2 hour classes will be held on Friday afternoons in the Spring term.

There will be material to read before each lecture, primarily essays and book chapters or excerpts, generally ~100 pages. The emphasis is on primary sources and original syntheses. There will be a final project.

Grading will be pass/fail.

Credits: 2

Class size will be capped at about 70.

The instructors

Blaise Agüera y Arcas leads a team at Google AI including researchers, engineers, and designers. Much of his team focuses on combinations of hardware, software and neural nets, including projects like the Clips camera, AI for Pixel phones, AIY Projects and EdgeTPU boards. They have also developed theory and technology for massively decentralized, privacy-preserving machine learning (Federated Learning), founded the Artists and Machine Intelligence program, and been advocates for fair and inclusive machine learning.

The following provide more background and relate to some of the themes of the course:

The last two are coauthored with Margaret Mitchell (Google AI) and Alex Todorov (Princeton).

Adrienne Fairhall is a Professor in the Department of Physiology and Biophysics and adjunct in the Departments of Physics and Applied Mathematics at the University of Washington. She obtained her Honors degree in theoretical physics from the Australian National University and a PhD in statistical physics from the Weizmann Institute of Science. She received her postdoctoral training at NEC Research Institute with Bill Bialek and at Princeton University with Michael J. Berry II. She is the co-director of the Computational Neuroscience Program at UW and, with Prof. Tom Daniel, co-directs the UW Institute for Neuroengineering. She has directed the MBL course Methods in Computational Neuroscience, is on the Advisory Board of the Allen Institute for Brain Science’s Mindscope project, and is a member of the BRAIN Initiative Working Group’s Neuroethics subcommittee. Her work focuses on dynamic neural computation, with a particular interest in the interplay between cellular and circuit dynamics and coding.

Reading list preview

Provisionally, readings will include selections from many of the following authors:

  • Charles Babbage and Ada Lovelace
  • Francis Bacon
  • Nils Barricelli
  • Gregory Bateson
  • Walter Benjamin
  • Jeremy Bentham
  • Nick Bostrom
  • Benjamin Bratton
  • Nancy Cartwright
  • Penny Chisholm
  • Patricia Churchland
  • Ta-Nehisi Coates
  • David Cope
  • Kate Crawford
  • Charles Darwin
  • Manuel DeLanda
  • Gilles Deleuze and Félix Guattari
  • Daniel Dennett
  • Jacques Derrida
  • George Dyson
  • Daniel Everett
  • Richard Feynman
  • Susan Fowler
  • Masha Gessen
  • Brooke Gladstone
  • Michael Graziano
  • Donna Haraway
  • Moritz Hardt
  • Graham Harman
  • Helen Hester
  • Jaron Lanier
  • Kai-Fu Lee
  • Warren McCulloch and Walter Pitts
  • Robert Merton
  • Timothy Morton
  • Alex Mordvintsev
  • Thomas Piketty
  • Steven Pinker
  • Kim Stanley Robinson
  • Jon Ronson
  • Douglas Rushkoff
  • Steven Shapin and Simon Schaffer
  • Marci Shore
  • Michael Taussig
  • Baratunde Thurston
  • Zeynep Tufekci
  • Alan Turing
  • Christopher Turner
  • Ellen Ullman
  • Paul Virilio
  • Judy Wajcman
  • Norbert Wiener
  • Anthony Wilden
  • Joanna Zylinska and Sarah Kember