Computer Ethics

Required:

• Read "Whose Life Should Your Car Save?" by Azim Shariff et al., 2016 (1 page)

  This is a high-level introduction to the moral machine project that also discusses how politics and business view ethical dilemmas, and that a tragedy of the commons-type scenario could undermine our common expectations for machine behavior.

• Read "Why the moral machine is a monster" by Abby Everett Jaques, 2019 (10 pages)

  This paper presents a fierce-yet-thoughtful critique of the Moral Machine experiment. It also highlights the importance of analyzing the structural implications of the problem: “In a slogan, the problem is that an algorithm isn’t a person, it’s a policy. And you don’t get policy right by just assuming that an answer that might be fine in an individual case will generalize. You have to look at the overall structure you create when you aggregate the individual transactions….The right question is what kind of world will I be creating if this is the rule. What will the patterns of advantage and disadvantage be if this preference is encoded in our autonomous vehicles. (pg 5-6)”

Optional:

• Read "The Trolley Problem" by Judith Jarvis Thomson, 1985 (21 pages)

  Judith Jarvis Thomson marches you through 21 pages of analyzing various hypothetical scenarios in order to expose the difficulty of using abstractions, and how a normative/ethical tenet may on the face seem straight forward, but really be undermined if one looks at the details. Makes it seem difficult to think that there are an exhaustive set of rules to encode which actions are permissible, not to mention moral.

• Read "Ethical Machines (published)" by Irving John Good, 1982 (5 pages)

  I.J. Good discusses problems with certain considerations one should make when thinking about whether and how a machine could exhibit “ethical” behavior.

• Read "Ethical Machines (unpublished)" by Irving John Good, 1980 (16 pages)

  The longer version is able to better parse the implications of the arguments in part because it’s more in depth due to its length. Notable in particular are the final consideration of a “synergistic relation with the boss” and the shakespeare reference.

Before Class:

• Daily assignment, due at 8pm the night before class:
  • Judge a few scenarios on the MIT Moral Machines website.
  • Summarize the readings.
  • What are the trade-offs of looking at individual ethics cases (such as in the “Trolley Problem” or “Moral Machines”) as compared to a ‘structural analysis’ as described by Jaques?

Required:

• Read "It's Not Privacy, and It's Not Fair" by Cynthia Dwork et al., 2013 (6 pages)

  This law review paper is the missing link between the concept of control and of privacy as represented by the (optional) Deluze piece and the Barocas piece, respectively.

• Read "Think You're Discreet Online? Think Again" by Zeynep Tufekci, 2019 (2 pages)

  How ought we make sense of questions such as privacy, classification, tracking, and surveillance in the era of big data and computational inference? Zeynep Tufeci asks us to consider these questions by looking at examples of the collective implications of a “privacy-comprised world”.

• Read "Big data's end run around procedural privacy protections" by Solon Barocas et al., 2014 (2 pages)

  Solon Barocas and Helen Nissenbaum, both well-known AI ethics scholars, consider ”why the increasingly common practice of vacuuming up innocuous bits of data may not be quite so innocent: who knows what inferences might be drawn on the basis of which bits?”

Optional:

• Read "Deleuze the Societies of Control"

  This video highlights some significant passages in “Postscript” and explains what’s going on by connecting it back to contemporary questions of control. Only the first 10 minutes actually cover the essay and the next 12 or so are on “commentary”, by posing relevant questions and extrapolating “Postscript”’s ideas into the future.

• Read "Postscript on the Societies of Control" by Gilles Deleuze, 1992 (4 pages)

  “just as the corporation replaces the factory, perpetual training tends to replace the school, and continuous control to replace the examination. Which is the surest way of delivering the school over to the corporation.” Deleuze considers the technologies of power, and what it means to be in a “control state”.

Before Class:

• Daily assignment, due at 8pm the night before class:

  Identify a new system or one we’ve discussed in class that makes decisions which affect people’s lives in some meaningful way. Describe it and then answer the following questions:

  • Does this system rely on data collection to make these decisions?
  • Where does this information come from?
  • What’s the consent model?
  • What questions related to individual privacy does it raise?
• Finish the course project, part one by tonight.

Required:

• Read "Physiognomy's New Clothes" by Blaise Aguera y Arcas et al., 2017 (16 pages)

  Can an algorithm detect criminality from your face alone? In this paper, the authors critique a 2016 paper that claims to be able to do so. By focusing on the historical context and current science, this article refutes the paper claims while providing a salient illustration of the danger of using Neural Networks inappropriately.

• Read introduction , background, and discussion (pg. 1 - 2; pg. 2 - 5; pg.10 - 14; 7 pages total) from "The Misgendering Machines: Trans/HCI Implications of Automatic Gender Recognition" by Os Keyes, 2018 (21 pages)

  This paper surveys the HCI papers to understand how researchers understand “gender” and the implications for how Automatic Gender Recognition systems are developed. As you’re reading this, think about who gets to decide what is “gender” when developing a product and how different operationalizations of this concept can impact how AGR systems are deployed and impact Trans people.

• Read "STEALING UR FEELINGS" by Noah Levenson, 2019

  Also consider watching the video

Optional:

• Read "Facial feature discovery for ethnicity recognition" by Cunrui Wang et al., 2019 (17 pages)

  This is another example of the ways facial recognition may err.

Before Class:

• Daily assignment, due at 8pm the night before class:
  • How does the field of physiognomy reinforce the power of some at the expense of others? Does the comparison to scientific racism seem reasonable?
  • Within the context of facial recognition systems, there are many incentives for increasing systems’ efficacy, such as to advance political or economic goals. How do these incentives influence:
    • The system’s ability to control and exert power?
    • Philosophies of human nature and essentialism?
    • Reputability of the science and facial recognition’s applications? In answering this question, pick a specific incentive of a political or economic nature and one of the above externalities.
  • What are possible blind-spots in our own conceptions of human traits?

Required:

• Read "The Automation Charade" by Astra Taylor, 2018 (5 pages)

  Astra Taylor urges: “We shouldn’t simply sit back, awestruck, awaiting the arrival of an artificially intelligent workforce. We must also reckon with the ideology of automation, and its attendant myth of human obsolescence.”

• Read page 5 onwards from "The Triple Revolution (original)" , 1964 (16 pages)

  This famous memo to President Lyndon B. Johnson was drafted by the Ad Hoc Committee of the Triple Revolution, comprised of notable social activists, scientists and technologists, among others. It warns that revolutions in social justice, automation, and weapon development, and that if urgent social and economic changes are not made “the nation will be thrown into unprecedented economic and social disorder”.

Optional:

• Read "Is this time different? The opportunities and challenges of artificial intelligence" by Jason Furman, 2016 (17 pages)

Before Class:

• Daily assignment, due at 8pm the night before class:

  These questions are structured to prepare you for the argument section of the course project.

  • Both of these readings make calls for action. Describe each. In doing so, consider: Who is the intended audience? Which stakeholders are considered?
  • Choose one rhetorical move which one of the authors made. What was it? Was it compelling? Examples include:
    • Use of narrative — what was the purpose of using storytelling?
    • Specific examples used by the author(s) and what made the example effective.
    • Enumerating concepts, facts, etc. — what did this do for the overall argument?
  • How might you incorporate a similar rhetorical move into your own writing (such as in the course project)?
• Finish the course project, part two by tonight.

Required:

• Read "It's the (Democracy-Poisoning) Golden Age of Free Speech" by Zeynep Tufekci, 2018 (4 pages)

  Zeynep Tufekci considers how power to censor functions on our oversaturated social networks, and the role of misinformation and the attention economy in this. The article provides striking clarity to issues we collectively face on this platform.

• Read chapter four (pg. 74 - 110; 36 pages total) from "Custodians of the internet: platforms, content moderation, and the hidden decisions that shape social media" by Tarleton Gillespie, 2018 (288 pages)

  Read the whole chapter, but, if you’re short on time, focus on the “automatic detection” section (pg. 97 to 110). Tarleton Gillespie, a social media scholar, establishes the groundwork for understanding social media platforms. In chapter 4, “Three Imperfect Solutions to the Problem of Scale”, Gillespie considers several models for content moderation, one of which is algorithmic “automatic detection” techniques.

Optional:

• Read "The Risk of Racial Bias in Hate Speech Detection" by Maarten Sap et al., 2019 (9 pages)

Before Class:

• Daily assignment, due at 8pm the night before class:

  In 150 - 300 words: Identify a system that attempts to alleviate some of the concerns raised about platforms. What does it do? What ideologies and values inform it? What potential abuses might be overlooked by this design?

Required:

• Read "Three Expensive Milliseconds" by Paul Krugman, 2014 (1 page)

  While reading this article, consider: who is today’s infrastructure for? What are some metrics being optimized for which have led to some (perhaps) unexpected consequences? How are these metrics and systems shaping our world?

• Read chapter 1 (pg. 13 - 36; 23 pages total) from "Pressed for time: The acceleration of life in digital capitalism" by Judy Wajcman, 2015

  Wajcman, whom we saw previously in ”Techne” and Belonging, investigates the demand for speed and efficiency in our current society, particularly as encouraged by technologies (like smartphones) and their creators (like silicon valley corporations).

• Read "Survival of the Richest" by Douglas Rushkoff, 2019 (3 pages)

  From Douglas Rushkoff, also the author of “The Internet is Acid, and America is Having a Bad Trip”: “Apocalypto – the intolerance for presentism leads us to fantasize a grand finale. “Preppers” stock their underground shelters while the mainstream ponders a zombie apocalypse, all yearning for a simpler life devoid of pings, by any means necessary. Leading scientists – even outspoken atheists – prove they are not immune to the same apocalyptic religiosity in their depictions of “the singularity” and “emergence”, through which human evolution will surrender to that of pure information.” This idea is investigated more deeply in his book “Present Shock”.

Before Class:

• Daily assignment, due at 8pm the night before class:

  Describe acceleration and at least one way it has shaped our world.

  Recall that in the “The Triple Revolution,” the authors observe:

  There is no question that cybernation does increase the potential for the provision of funds to neglected public sectors. Nor is there any question that cyberna­tion would make possible the abolition of poverty at home and abroad. But the in­industrial system does not possess any adequate mechanisms to permit these poten­tials to become realities. The industrial system was designed to produce an ever-increasing quantity of goods as efficiently as possible, and it was assumed that the distribution of the power to purchase these goods would occur almost automati­cally. The continuance of the income-through-jobs link as the only major mechanism for distributing effective demand—for granting the right to consume—now acts as the main brake on the almost unlimited capacity of a cybernated productive system.

  In what ways has the current structure of our regulatory and political framework allowed for cybernetic systems (such as social media platforms and the stock market) to shape our world? Is this the world that the systems’ creators set out to create? If not, list some unintended consequences. What are the mechanisms that shape these systems? Who gets to decide?

Required:

• Read foreward and introduction (pg. xi - xxvi; pg. 1 - 5; 23 pages total) from "Hope in the dark: Untold histories, wild possibilities" by Rebecca Solnit, 2016

  Solnit, a writer and activist, reflects on our desire for social, cultural, or political change given the appearance that we have not arrived there (considering issues from global warming, to human rights abuse). Originally responding to the war in Iraq, she explores how news cycles and our personal narratives frame these issues and makes the case for hope nonetheless–“tiny and temporary victories”

Before Class:

• Daily assignment, due at 8pm the night before class:
  • Do you feel able to change outcomes of how tech affects society?
  • What’s an idea from this course that every UW CSE student ought to understand?
• Finish the course project, part three by tonight.