A banner image of the Responsible Computer Science Challenge


After four years and a pandemic, the first cohort of grantees from the Responsible Computer Science Challenge finally came together on November 4th to share their accomplishments and learnings.

This cohort’s success has been truly incredible to witness: collectively, they have created 100+ distinct classes for more than 15,000 students, and their efforts continue to spread to new institutions and contexts across the world. In the midst of drastic upheaval at Twitter, dystopian visions of the metaverse, and the spread of misinformation on social media platforms, educating a new generation of students who can critically examine the social implications of technology has never been more pressing.

Following Mozilla’s Theory of Change for Trustworthy AI, the Responsible Computer Science Challenge supports a global network of educators and students to catalyze a broader culture shift in the tech industry. Aside from the huge volume of resources that the grantees have created – which spans lecture materials, software, podcasts, YouTube videos, and case studies – grantees and contributors to the Challenge also created the Teaching Responsible Computing Playbook, which provides best practices for educating a new generation of students to think holistically about the design of technologies. A persistent theme throughout each of the grantees’ remarks was the strong emphasis on meaningfully integrating these concepts into core computer science courses so that ethics is at the forefront of technology development and design, and never an afterthought.

Below, find a summary of each grantee’s accomplishments and learnings:

___________

At Allegheny College, a team consisting of Oliver Bonham-Carter, Janyl Jumadinova, Greg Kapfhammer, and Doug Luman taught 30+ courses with integrated ethics content reaching over 700 students, all while launching a podcast and video series called ethiCS that highlighted ethical issues in computer science. Working with students, this team also designed, tested, and released open-source tools to facilitate ethically-inclined activities and assignments in existing computer science courses, along with a Proactive Programmers website that helps students ethically implement software projects on data abstraction and discrete structures. As a result of these courses, undergraduate students in this program created three open-source software tools: AFLuent, TaDa!, and CommitCanvas.

___________

The sociologist and computer scientist duo Colleen Greer and Marty J. Wolf at Bemidji State University took a different tack: while they also created new teaching modules and a new course, they developed criteria and methodology for practically evaluating the efficacy of new curricula on ethics, which is documented on their project website. This stunning resource includes ready-to-go teaching modules, over 40 module evaluations conducted by a panel of experts, and workshop materials that faculty can use to bring together colleagues from computer science and the humanities and social sciences. These efforts also culminated in a slew of conference and paper presentations at Ethicomp, the Midwest Sociological Society, and the Special Interest Group on Computer Science Education (SIGCSE).

___________

At Colby and Bowdoin Colleges in Maine, an interdisciplinary team of faculty in cinema studies (Allison Cooper), computer science (Stacy Doore), and digital studies (Fernando Nascimento) created an incredible online resource called Computing Ethics Narratives (CEN). This includes an impressive slew of papers and presentations at Ethicomp 2022, the American Philosophical Association, and the premier conference in computer science education, SIGCSE in 2021, 2022, and 2023. Students at both colleges have also collaborated with faculty members to pilot new modules and diversify the narrative collection in the repository, which includes a student-produced video that introduces the project on the repository website. Since the start of the project, Colby’s Computing Ethics course is now one of the two required courses in the College’s new AI concentration while Bowdoin is building on the Mozilla-funded work with a National Humanities Center and Google Foundation grant. To date, the CEN project includes 14 CEN modules which has reached 464 students in nine courses taught by computer science faculty.

Beyond their institutions, the Colby-Bowdoin team has also made great strides in creating and sustaining a global community of responsible computing advocates, which includes developing a course for Pan-Atlantic University in Nigeria, giving invited talks at the Harvard / MIT Embedded EthiCS conference, Louisiana State University, and Scripps College. Stacy Doore has also joined the ACM Joint Task Force on Computing Ethics (which includes RCS colleague Atri Rudra) and is the co-editor for the ACM Engage CSEdu Special Issue on Responsible Computing (with RCS colleagues Ellen Zegura and Benjamin R. Shapiro).

___________

The Colby-Bowdoin collaboration is not the only cross-institutional team. In a star-studded project created by computer scientists Sorelle Friedler (Haverford College), Suresh Venkatasubramaniam (University of Utah and Brown University), Seny Kamara (Brown), and Kathi Fisler (Brown), the Responsible Problem Solving project features slides, assignments, and pedagogical modules that are core components of standard data structures and algorithms curriculum. From criminal justice to the environment, these modules provide concrete templates for faculty to begin writing problem sets and lecture notes. In 2021, both Friedler and Venkatasubramaniam modeled best practices in responsible computer science by taking new positions in the White House Office of Science and Technology Policy advising on machine learning bias and algorithmic accountability.

At Haverford, this work has led to a department-wide restructuring and overhaul of the core data structures and algorithms curriculum and expanded to include the entire introductory sequence, a new 200-level required course, related courses at Bryn Mawr College, and a department-wide commitment to integrating responsible computer science ideas throughout the curriculum. The materials developed by this team have also been transferred to and adopted by other faculty in the department. At Brown, this project also expanded the computer science department’s socially responsible computing effort in 15 courses by supporting a specialized program for undergraduate teaching assistants. Since then, faculty have begun to articulate a multi-year curriculum that creates a broader arc for students to follow as they progress through the major.

___________

Interdisciplinary efforts featuring faculty from computer science, journalism, psychology, political science, anthropology, philosophy, and design yielded interesting projects that underscore the importance of a holistic approach to computer science education. At Columbia University (led by Augustin Chaintreau), a network analysis course with transformations in ethical content is now in the process of becoming a core class in the major, and the resulting publications from this work in Nature Scientific Reports, ACM Special Interest Group for Computer Performance Evaluation (SIGMETRICS), and ACM Economics and Computation. These publications have been ranked among the 10 most downloaded articles on the Social Studies Research Network (SSRN).

___________

An impressive team of faculty members at Georgetown University across the disciplines of computer science, philosophy, and design collaboratively designed and co-delivered integrative ethics exercises for three undergraduate computer science courses. The collaboration between Georgetown’s Ethics Lab and the computer science department led to the development of three new undergraduate programs in Tech, Ethics, and Society. Nearly 400 students took these courses, and their exercise on inclusive human-computer interaction is now publicly available on the EThics Lab website. At Georgia Tech and Georgia State, an interdisciplinary team created, piloted, and expanded several role play exercises as a way for students to understand the ethical implications of their work. These scenarios included the deployment of self-driving buses and the use of AI assessment in an undergraduate admissions process. Each of these policy simulations include documents, talking points, and exercises for students to engage with one another. They also include “train the trainers” materials such that teaching assistants could also use many of these concepts in other courses they were teaching. Since then, this team led by Ellen Zegura has also piloted a community network platform for TAs and translated the work from the Mozilla network into an education school context so that teachers are similarly prepared to teach computer science ethics in both K-12 and university contexts.

___________

The team at Harvard University led by Barbara Grosz pioneered the program Embedded EthiCS, which has created over 100 modules in over 40 distinct computer science courses. Embedded EthiCS now includes the majority of undergraduate CS courses at Harvard in any given term. Aside from this huge volume of curriculum development, the program also began testing a “co-piloting approach” – for CS instructors to take the lead on well-developed modules with the assistance of a postdoc. As the program continues to grow, the team at Harvard is also engaging in new outreach and collaboration and have since formed a Boston-area consortium called teχnē (Technology Ethics eXchange of the NorthEast) which includes an annual conference on research and teaching in the ethics of technology. While the program is at a state of steady transitioning towards a more stable state with regards to staffing and leadership, the Embedded EthiCS model has already been adopted or studied in multiple different schools: the University of Toronto, Stanford, Technion, and Tufts, among others.

___________

Miami Dade College is currently developing an interdisciplinary certificate between the philosophy department with an AI ethics course at the associate’s degree level. On an international level, the team is also structuring student research opportunities to learn how to recycle and repair computers through an open-source summer program in Belize.

The program at Northeastern University successfully pioneered their model of computer science assessment modules, which led to two permanent hires to continue delivering the modules. Since the end of the grant, the team led by Christo Wilson has acquired internal funding to expand their work to Northeastern’s partner campuses in the UK and California. More of this success can be documented in their initial assessment of the module outcomes.

___________

Over the course of both stages of RCS, faculty at Santa Clara University reached about 600 students directly, and they have consolidated their existing teaching resources into a new page titled “Embedding Ethics into Computing Curricula: Resources and Suggestions.” While many faculty members had long been innovating in the area, others were eager but did not feel like they had the expertise necessary to do so: these faculty acted as the go-between in order to facilitate collective learning and to identify different learning outcomes across these courses.

___________

Faculty at the University of Buffalo (Matthew Bolton, Kimberly Boulden, Varun Chandola, Adrienne Decker, Garegin Grigoryan, Matthew Hurtz, Andrew Hughes, Lindsay Hunter, Kenneth Joseph, Oliver Kennedy, Jonathan Manes, Dalia Muller, Atri Rudra, Maria Rodriguez, Mark Shepard, Sama Waham, and Jennifer Winikus) created responsible computing assignments and modules across five undergraduate computer science courses, started a new course titled Machine Learning and Society, and pioneered a module titled the “Impossible Project: Making Computing Anti-Racist,” which is based off of historian Dalia Muller’s radical Impossible Project framework (video, news coverage of finale event).

___________

The team at the University of California, Berkeley had an outsized impact on the computer science curricula: To date, they have reached about 7,000 students a year (20,000 students total) across three of the largest classes at Berkeley. They have conducted training for student TA teams (80 students/year) and cultivated a student team of curriculum developers. The project has gained an institutional platform and gained advocates in all three major programs (data science, computer science, and statistics), which has supported the development of courses on anti-racism in computing/statistics and the social implications of computer technology. The curricula that these faculty have developed have been integrated across the curricula (e.g. classes such as “data and justice,” “foundations of data science”) and resulted in publications about ethical thinking and data science practice, as well as a “human context and ethics” toolkit for use across courses.

___________

At the University of Colorado, Boulder, the team led by Casey Fiesler created and implemented introductory programming assignments about personalized ads, college admissions algorithms, content moderation, and password security. They also iterated on and evaluated the “Black Mirror Writers Room” teaching exercise, and conducted an interview study to evaluate its efficacy. They also researched instructor attitudes towards ethics integration into computing classes. They found that while instructors were generally positive about ethics as a component of computing education, there were often specific barriers to doing so, and they require institutional support structures (e.g., community with which to talk about ethics). Throughout the course of the grant, Fiesler has blogged extensively about the importance of responsible computing, particularly about ethics integration, what we teach when we teach ethics, what counts as computer science, and ethical speculation in the CS classroom. She also has a YouTube video essay about the case for ethics integration.

___________

At the University of Maryland-Baltimore County, instructors created and shared introductory computing and programming instructional materials, which included lectures, case studies, and final class projects. Their strength throughout the program was their extensive research on computing education. In one study on peer-led ethics learning, they evaluated the impact of introducing first-year computer science (CS) students to ethical thinking about the social justice impacts of data collection, tracking, bias, internet privacy, and competitive “real world” system design and critique activities. While basic content was consistent for all, one group was involved throughout the course in peer discussions designed to foster greater engagement, with the anticipation that this would enable students to reach new levels of sensitivity through peer-to-peer interaction. The research showed that students are sensitive to the technology-related risks and vulnerabilities encountered by individuals based on race, gender, and, to some extent, age, but they struggle to assess who is responsible for these risks, what to do about bias in technology design, and how to mitigate harms for individuals whom they perceive to be vulnerable, furthering the argument for an integrated ethics curriculum. The paper further explores the value of formal peer-led discussion to evolve social justice thinking with a focus on identity, though note that opportunities for any group discussion are meaningful to students’ thinking about social justice. Over the longer term, students tend to recall and apply ethics that are closely related to their identity, suggesting that empathy has limits.

___________

At Washington University in St. Louis, instructors created canvas lessons on fair division algorithms, and a new course titled “Data Science Playground” that features faculty presenting case studies at the intersection of data science and ethics / responsibility. In the university’s studio sessions, students formed groups of five to unpack responsible CS together across an entire semester. And prior to COVID, instructors successfully deployed a module in the introductory CS course to study algorithmic fairness through cake cutting and algorithmic unfairness by a biased auction scheme. Further, instructors in economics, political science, chemistry, environmental studies, English, and history spoke about how they use data science in their research, and highlighting the ethics and responsibility issues associated with that work.

Next steps

Although the grant-making period for this cohort has ended, their essential work in the classroom will continue in the years ahead. Meanwhile, the students who engage with these curricula will take valuable lessons from the classroom into the workforce.

The Responsible Computer Science Challenge is also growing into an international initiative. In August, Mozilla announced that with the help of USAID, the project will expand to higher education initiatives in Kenya and India.

We’re excited to see this work continue and evolve, both in the U.S. and in the challenge’s new terrain.

—

The Responsible Computer Science Challenge is funded by the Omidyar Network, Mozilla, Schmidt Futures,Craig Newmark Philanthropies, USAID, the Andrew W. Mellon Foundation, and the Rockefeller Brothers Fund.