Diversity, Inclusion and Leadership in Tech
This is a reading seminar on issues of diversity and inclusion, focusing on how to take leadership in creating more equitable and just communities in tech.
Spring 2022 Diversity, Inclusion and Leadership in Tech Reading Group
This is a reading group for IACS students and affliates on issues of diversity and inclusion, focusing on how to take leadership in creating more equitable and just communities in tech. The reading group is supported and facilitated by members of the IACS Graduate Advisory Committee as well as by IACS faculty and staff.
This reading group is a not-for-credit continuation of the Fall Diversity, Inclusion and Leadership Seminar. The reading group will be bimonthly and discuss papers on topics that are chosen by both group facilitators as well as participants.
We welcome folks of all backgrounds in the wider IACS community to sign-up for the reading group!
Location and Time
We will meet bi-monthly on Fridays from 3.30pm-4.30pm in SEC 6.301.
Erin Erhart will host a DIB coffee hour on alternate Fridays (3.30pm-4.30pm) in their office SEC 1.312.09 (see Schedule).
Prerequisite Readings:
In order to establish a foundation of shared language and references, we ask that participants complete the following list of reading prior to our first group meeting. These readings are selected from the syllabus of the DIL Seminar.
- Five Years of Tech Diversity Reports—and Little Progress
- WHY IS SILICON VALLEY SO AWFUL TO WOMEN?
- Introduction and Chapter 2 from Race After Technology
- Chapters 1 and 8 from Invisible Women
- Data Science as Political Action: Grounding Data Science in a Politics of Justice
- “Those invisible barriers are real”: The Progression of First-Generation Students Through Doctoral Education
- Fourteen Recommendations to Create a More Inclusive Environment for LGBTQ+ Individuals in Academic Biology
- Awareness in Practice: Tensions in Access to Sensitive Attribute Data for Antidiscrimination
- Introduction and Chapter 1 of Ghost Work
- Algorithms are Not Neutral: Bias in Collaborative Filtering
Schedule
February 4th, 3.30pm-4.30pm Introductions
- Reading: Prerequisite Readings
- Suggested discussion questions:
- From the readings, what is your sense of the current landscape of diversity in tech? Does this align with your personal experiences in academia/industry?
- From the readings, how might one argue for increasing diversity in tech – i.e. why is diversity in tech important, what are some of the negative consequences of the lack of diversity. Does this align with your personal experiences in academia/industry?
- From the readings, what are the obstacles to diversifying tech? Where are the leadership opportunities for change in the wider tech community? Where are they at IACS or Harvard?
- What would you like to get out of this reading group? What topics/issues would you like to explore?
February 11th, 3.30pm-4.30pm DIB Coffee Hour with Erin Erhart
February 18th, 3.30pm-4.30pm Building Trust Across Differences & Tackling Difficult Subjects
- Reading:
- Suggested discussion questions:
- How would you explain the differences between cultural humility and cultural competence? Why is cultural humility more beneficial for understanding diversity?
- What role might cultural humility play in navigating difficult conversations (e.g. about race)? In particular, which concrete suggestions in Guide for Navigating Difficult Conversations on Race might exemplify cultural humility.
- The guidelines provided in Guide for Navigating Difficult Conversations on Race is for navigating conversations about race. Which aspects of these guidelines are generalizable to other conversations about sensitive topics or to situations where we need to build trust across differences?
- Last week, we brought up the question of how to have conversations about issues that acknowledge harm (to individuals, communities etc). In what ways might we be neglecting to address harm in the current scientfic/professional discourse on data science? In Situating the Scientist: Creating Inclusive Science Communication Through Equity Framing and Environmental Justice, can equity framing provide a way to acknowledge and address harm?
- How does framework proposed in Situating the Scientist: Creating Inclusive Science Communication Through Equity Framing and Environmental Justice relate to cultural humility and the principles described in Guide for Navigating Difficult Conversations on Race?
February 25th, 3.30pm-4.30pm DIB Coffee Hour with Erin Erhart
March 4th, 3.30pm-4.30pm Situating the Scientist
- Reading:
- Suggested discussion questions:
- Why is problem framing important in determining the broader impact of tech? Can you give a real-life example where, depending on who gets to frame the problem and how the problem is framed, different problem framings result in drastically different technical solutions with different broader social impacts?
- What is a Solutionism mindset and how can it stand in the way of addressing wider social issues?
- Thinking back on your data science education, how have technical problems you’ve encountered in classes been framed? Have these framings made explicit their social/polical/cultural assumptions? Can you think of instances where technical problem framings can lead to technical solutions with potentially undesirable broader soical impacts?
- In Chapter 3: On Rational, Scientific, Objective Viewpoints from Mythical, Imaginary, Impossible Standpoints, what are the major critiques of our current practices of presenting data, i.e. how can “objective”, “clinical” presentations of data encode “subjective” values and narratives? How does this relate to the recommendations for scientific communication in last week’s reading? Specifically, in Situating the Scientist: Creating Inclusive Science Communication Through Equity Framing and Environmental Justice, the authors present a case for why the emphasis of “overpopulation” is problemmatic in context of environmentalism messaging, how does their analysis compare with the critique of standard data presentations in Chapter 3: On Rational, Scientific, Objective Viewpoints from Mythical, Imaginary, Impossible Standpoints?
- In Critical Race Theory for HCI, 9 personal stories are featured. What is the purpose of featuring “subjective” personal narratives? What are the functions of personal narratives that cannot be replaced by standard presentations of data (e.g. summary statistics)?
- In Critical Race Theory for HCI, in what ways do the personal stories highlight the ordinariness of race (i.e. how do we concretely instantiate the idea that race is ordinary)? Which pitfalls, commonly experienced by researchers who are engaging with issues involving race, are highlighted in the personal narratives?
- In last week’s readings, we encountered the idea that “race is a social construct”. How can we understand this idea without concluding that we should ignore the role that race plays in our socio-technical systems (risking the erasure of the real experiences of racism)? How can we, as scientists, meaningfully engage with race and other identities?
March 25th, 3.30pm-4.30pm Case Study I: The Data Nutrient Project
In this session, we look at ways that we can instantiate princples/frameworks from our readings to evaluate the broader social impact of technology.
- Motivation: We have all heard the phrase “garbage in, garbage out”, referring to the fact that models built on unreliable data will be unreliable for down-stream tasks. Now, we want to study concrete ways in which undesirable properties of data may cause real-life societal harm through their effects on machine learning models. The Data Nutrient Project is an example of this type of endeavour. First, familiarize yourself with the mission of the Data Nutrient Project.
- Dataset: As a case study, let’s generate a data Nutrient label for the NYC Notice of Property Values dataset.
- Activity: For the NYC Notice of Property Values dataset, we want to answer the the following questions about the dataset. Furthermore, we want to ask why each question is important for gauging downstream social impacts of the dataset.
- Reference Data Nutrient Label: For reference, compare your data Nutrient label to the one generated by the Data Nutrient Project. Do your answers agree with theirs?
April 8th, 3.30pm-4.30pm Roles of Data Science in Social Change
This week we examine another case study of the broader socal impact of technology. In particular, we examine frameworks and examples for applying data science for social change and advocacy.
- Reading:
- Suggested discussion questions:
- Roles for Computing in Social Change highlights a number of categories of ways that computing can contribute to social change. In which category, if any, does the work described in A large-scale analysis of racial disparities in police stops across the United States fall? Can you think of some recent examples of “AI for social good” projects and map them to the categories outlined in Roles for Computing in Social Change?
- What are some potential draw-backs and blind-spots of the study in A large-scale analysis of racial disparities in police stops across the United States? What are some limitations on the generalizability of the insights/conclusions in the paper?
- What are the implementational bottlenecks for each category of computing application in Roles for Computing in Social Change(e.g. what are the political, economic, cultural challenges that must be overcome in order to achieve social change via tech in each category)?
- Do you agree with the taxonomy of computing applications in Roles for Computing in Social Change?
April 22th, 3.30pm-5pm Data Labeling Party with the Data Nutrition Project
This week the Harvard Data Nutrition Project is visiting our reading group. We will engage in another short data labeling exercise and discuss research in ML transparency, fairness and policy.