Leveraging data equity and putting “Tech for Good” into practice
By Alexi Orchard, CML PhD Candidate and Lab Manager
In recent years, most people in a computing or technology-related profession have likely been exposed to “tech for good,” “responsible innovation,” “computing for social good,” or one of the many other buzzwords and phrases that advocate designing technology for the benefit of humanity and the environment.
As a PhD candidate and teacher of engineering ethics and communication at the University of Waterloo, my primary goal is to prepare students to think critically and design responsibly in their future careers. Students are often receptive to “tech for good,” as seen in manifestos and principles like the Canadian Good AI Coalition’s Tech for Good Declaration, Microsoft’s Responsible AI Standards, or Meta’s Responsible Innovation Principles.
However, the vague, corporate-speak phrasing of “leave no one behind” (Tech for Good) or “provide controls that matter” (Meta) is not readily actionable for students, or most industry professionals for that matter.
Fortunately, more resources are becoming available to help data practitioners make ethical and equitable decisions in their workflow. LA Tech4Good, with their flagship Leading Equitable Data Practices workshop, is one of the leaders in this space. After participating in the November 2022 cohort of the workshop, I was excited to share some of the lessons learned in a workshop with some of my third-year undergraduate engineering students.
Systems Design Engineering Course 362
When designing equity-based outcomes, Catherine D’Ignazio and Lauren Klein (authors of Data Feminism) and Sasha Costanza-Chock (author of Design Justice) provide fundamental inspiration for any aspiring (or experienced) data practitioner. For example, Engineering students are excellent at math, but sometimes lean on the assumption that numbers are purely objective. The fourth chapter of Data Feminism, “What Gets Counted Counts”, flips the script by challenging normative systems of counting and binary classification.
This classic Facebook sign-up shows the binary gender options (female and male) and is still all too common!
After our Engineering class dug into these fundamentals, we also discussed the Gates Foundation’s “Reinvent the Toilet Challenge,” a $200 million project for worldwide project teams to design affordable sanitation solutions in developing countries. It turns out that making a new toilet does not account for the whole context: sanitation experts from the developing countries concluded that their communities faced issues of “access, social mobilization, and ongoing maintenance” that would make the project’s innovations unfeasible, reports Costanza-Chock. The moral of the story is aptly articulated by Kimberly Seals Allers, “Whatever the question, the answer is in the community.”
With this case study in mind, I presented the class with their own design challenge and sent them off into breakout groups:
Challenge: The Region of Waterloo has requested a report on the public’s use of Waterloo Park. Write a report through an equity lens that will inform park service, maintenance, and future development.
In the first breakout time, students identified the relevant stakeholders and discussed what metrics and methods of data collection would best fulfill the challenge. This first step came naturally to the students. We’re able to envision the high-level functions of a public park and apply different methods to describe the activities and experiences of participants in this space (see Figure 1); however, much like installing high-tech toilets in a developing country, context and community engagement are essential.
As noted by D’Ignazio and Klein, “People in a community know its problems intimately, and they know which phenomena go uncounted, underreported, or neglected by institutions in power (or, conversely, who is overly surveilled by institutions in power).” The metrics we decide to include (or exclude) and our proximity to the design, problem spaces, and communities will have a tremendous impact on what can be accomplished.
In the second breakout time, students were prompted to identify how structural problems, including colonialism, classism, and ableism (among many others), might influence their metrics and methods for data collection. We reviewed Gebru et al.’s Datasheets for Datasets, asking questions about “invisible” groups or categories, or metrics for addressing equity specifically.
Equity can be considered in every stage of a project
One of our main takeaways was that equity can be considered in every stage of a project – from the mobile accessibility of a survey to the inclusivity of public signage. Cyclists, children, and unhoused people all have significant stakes in the functioning of a public park. Public parks also provide important resources to people in low-income neighborhoods.
These stakes are heightened in areas with significant heat waves, like those seen in Southern California in recent years.
According to the Los Angeles’s newly appointed Chief Heat Officer Marta Segura, “neighborhoods without trees, shade, and open space suffer four times the number of hospitalizations and premature deaths than areas with a greener, more climate-adapted infrastructure.” In 2022, Segura’s team launched Cool Spots LA, a web application showing residents where they can seek relief during major heat events. The team incorporated equity into this project by gathering data on augmented cooling centers such as libraries, recreation centers, and senior centers, and cooling spots such as shade, bus shelters, and hydration stations. Equitable access to information and physical resources has been key to this implementation.
We – engineers, designers, and educators – must listen to and collaborate with the communities that our work impacts most. Then, eventually, we can accomplish a sense of “tech for good.”
By embedding some of the LA Tech4Good resources into this workshop, I was able to help students envision a process for applying their technical knowledge to the equity-related considerations of a shared public space. It is an empowering experience when students can unlock their skills to improve social outcomes like we discussed in this workshop. Students followed up with questions about applying these concepts to their capstone projects; I eagerly look forward to seeing their results.
I was left feeling rewarded and reminded of this: We – engineers, designers, and educators – must listen to and collaborate with the communities that our work impacts most. Then, eventually, we can accomplish a sense of “tech for good.”
This article was written in collaboration with and cross-posted by LA Tech4Good. Original post can be found here.