Governing Digital Technology

There has been a vigorous, decades-long debate around the governance of digital technology – one conducted variously in the public sphere, in academia, and among regulators and legislators – with a discernible impact on the development and use of consumer-facing technologies. These technologies include the Internet and applications built on top of it; AI and machine learning; and the digital platforms that began as individual apps but then became themselves a more comprehensive, at times totalizing, infrastructure of ubiquitous sensors, data, and user analysis. 

This course will contextualize today’s technology-driven policy challenges in terms of some of the complex technical, legal, and political arcs of digital governance. Throughout the semester we will trace a broad range of urgent, seemingly novel governance problems back towards their roots for the purpose of charting potential paths forward. We will take a long view on questions of technology policy, looking for the limitations and opportunities confronting both regulators and private-sector decision makers – and the users of these technologies. 

5 Beyond Internet Governance 5 Beyond Internet Governance

5.1 Week 11: Algorithmic Accountability (4/10 & 4/11) 5.1 Week 11: Algorithmic Accountability (4/10 & 4/11)

5.1.1 AI and Bias 5.1.1 AI and Bias

5.1.1.1. Algorithms of Oppression by Safiya Noble, New York University Press (2018) [Read Introduction and Chapter 1: “A Society, Searching.”].

62 PAGES

Algorithms of Oppression by Safiya Noble, New York University Press (2018) [Read Introduction and Chapter 1: “A Society, Searching.”].

5.1.1.2. “Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification” by Joy Buolamwini and Timnit Gebru, Conference of Fairness, Accountability and Transparency (2018)

12 PAGES

“Gender Shades: Intersectional Accuracy Disparities in Commercial Gender Classification” by Joy Buolamwini and Timnit Gebru, Conference of Fairness, Accountability and Transparency (2018)

5.1.1.3. State of Wisconsin v. Eric Loomis (2016), paragraphs 11-28 and 67-74

10 PAGES

State of Wisconsin v. Eric Loomis (2016), paragraphs 11-28 and 67-74

5.1.1.4. “Machine Bias” by Jeff Larson, Surya Mattu, Lauren Kirchner and Julia Angwin, ProPublica (2016)

15 PAGES

“Machine Bias” by Jeff Larson, Surya Mattu, Lauren Kirchner and Julia Angwin, ProPublica (2016)

5.1.1.5. “COMPAS Risk Scales: Demonstrating Accuracy Equity and Predictive Parity” by Northpointe, Inc (Northpointe, 2016) [Read Executive Summary, Introduction, Conclusion; skim Results]

4 PAGES

“COMPAS Risk Scales: Demonstrating Accuracy Equity and Predictive Parity” by Northpointe, Inc (Northpointe, 2016) [Read Executive Summary, Introduction, Conclusion; skim Results]

5.1.1.6. “Algorithmic Recommendations and Human Discretion” by Victoria Angelova, Will Dobbie, and Crystal Yang. (2022). Publication forthcoming. [Read Section I]

3 PAGES

“Algorithmic Recommendations and Human Discretion” by Victoria Angelova, Will Dobbie, and Crystal Yang. (2022). Publication forthcoming. [Read Section I]

5.2 Week 12: Machine Learning and Intellectual Debt (4/17 & 4/18) 5.2 Week 12: Machine Learning and Intellectual Debt (4/17 & 4/18)

5.2.1 Can AI governance learn anything from Internet governance? 5.2.1 Can AI governance learn anything from Internet governance?

5.2.1.1. “Artificial Intelligence — The Revolution Hasn’t Happened Yet” by Michael I. Jordan, Harvard Data Science Review (2019)

10 PAGES

“Artificial Intelligence — The Revolution Hasn’t Happened Yet” by Michael I. Jordan, Harvard Data Science Review (2019)

5.2.1.2. “The Bitter Lesson” by Rich Sutton (2019)

1 PAGE

“The Bitter Lesson” by Rich Sutton (2019)

5.2.1.3. “The Hidden Costs of Automated Thinking” by Jonathan Zittrain, The New Yorker (2019)

4 PAGES

“The Hidden Costs of Automated Thinking” by Jonathan Zittrain, The New Yorker (2019)

5.2.1.4. ”Artificial Intelligence’s ‘Black Box’ Is Nothing to Fear” by Vijay Pande, The New York Times (2018)

2 PAGES

”Artificial Intelligence’s ‘Black Box’ Is Nothing to Fear” by Vijay Pande, The New York Times (2018)

5.2.1.5. “Stop Doing Explainable ML” by Cynthia Rudin, Talk from “Statistics at a Crossroads: Challenges and Opportunities in the Data Science Era” (2018) 10 MINUTES Print version at https://rdcu.be/bBCPd

“Stop Doing Explainable ML” by Cynthia Rudin, Talk from “Statistics at a Crossroads: Challenges and Opportunities in the Data Science Era” (2018) 10 MINUTES Print version at https://rdcu.be/bBCPd

5.2.1.6. “AI models: what alignment do they have? Do they have alignment? Let’s find out!” by Jon Evans, Gradient Ascendent (2022)

8 PAGES

“AI models: what alignment do they have? Do they have alignment? Let’s find out!” by Jon Evans, Gradient Ascendent (2022)