The Responsible Tech Movement in Academia
An overview of the panel conversation we held at the recent Responsible Tech University Summit with notes by Francesca Cavuoti
Panel conversation with Saima Akhtar (Associate Director, Vagelos Computational Science Center at Barnard College), Steven Kelts (Lecturer, Princeton University), and Vasiliki Bednar (ED, McMaster University MPP in Digital Society) with moderator Chhavi Chauhan (Director for Scientific Outreach at the American Society for Investigative Pathology).
Dr. Saima Akhtar is the Associate Director of the Vagelos Computational Science Center (CSC) at Barnard College. She is a computational social scientist with a background in architecture and software engineering. Prior to joining Barnard, Saima was a postdoctoral associate in the Yale University Department of Computer Science, where she managed digital cultural heritage preservation projects between the fields of computer science and architecture.
Steven Kelts is a long-time ethics educator. He has twenty years of experience working with undergraduates, including in residential education environments and intensive, selective seminars. His research is on the history and uses of market ideas, including theories of the firm and corporate organization. He consults in the private sector with companies looking to synergize their market value with their ethical values, and to develop curricula to help their employees navigate ethical pitfalls in their organizational culture. https://gradfutures.princeton.edu/grad-stories/steven-kelts
Vass Bednar is a public policy entrepreneur working at the intersection of technology and public policy. She is an interdisciplinary wonk focused on ensuring that we have the regulatory structures we need to embrace the future of work and new ways of living. As an enthusiastic and perpetual student of the policymaking process, she has held leadership roles at Delphia, Airbnb, Queen's Park, the City of Toronto, and University of Toronto. Vass is recognized as a creative, data-driven thinker and was the Chair of the federal government's Expert Panel on Youth Employment. A graduate of McMaster University's Arts & Science Program, Vass holds her Master of Public Policy (MPP) from the University of Toronto and successfully completed Action Canada and Civic Action DiverseCity Fellowships. Passionate about public dialogue, she was also the co-host of "Detangled," a weekly pop-culture and public policy radio show and podcast that ran from 2016-2018. She currently writes a newsletter about Canadian startups and public policy called "regs to riches" and was recently recognized as an outstanding alum with a McMaster "Arch" award.
Dr. Chhavi Chauhan works as Director for Scientific Outreach at the American Society for Investigative Pathology and Director of the Continuing Medical Education (CME) Program at the Journal of Molecular Diagnostics. She is one of the leaders of the Women in AI Ethics Collective and an expert at the AI Policy Exchange. She is a biomedical researcher, expert scholarly communicator, and a sought after mentor in the fields of scientific research, scholarly publishing, and AI Ethics, especially for women and minorities. She is a thought leader, a renowned international speaker, and a strong advocate for equitable and accessible healthcare. She sits at the intersection of scientific research, scholarly communications, and AI Ethics in Healthcare. Her vision is to provide equitable personalized healthcare to all, beyond geographies, and despite socioeconomic disparities.
NOTES by Francesca Cavuoti
Question: pipeline issue, do you think we have a pipeline problem to integrate responsible tech in academia?
Bednar: responsible tech framework is less sophisticated in Canada, what roles should exist and how do you embed them in small and medium size companies?
Akhtar: Pedagogy could help, there is a need to think outside of computer science as a discipline. With the pandemic, we are starting to gather virtual spaces, and pay more attention to the issues of how marginalized and disadvantaged communities are impacted by virtual spaces.
Kelts: working with graduate students and PhD students, who have specialized training. Because of that, conversations about how the field calls for anthropologists, computer scientists, public policy experts to talk of where tech is going. Tech touches all aspects of our life, therefore we need experts to understand how to regulate and govern all those aspects.
In case of unprecedented events, we need to adapt and react to current events. Based on opening statements, do you see some model or pedagogical concern of what we are currently teaching?
Kelts: many undergrad ethics courses focus on textual analysis. But its also important to get experience encountering ethical dilemmas as you might find them in the workplace. Introduced workflow wednesday, talking about engineering workflows, development processes, and where we should raise ethical dilemmas.
Akhtar: center for computational science in historically women's liberal arts college. Sometimes it's not a natural fit, but the latest literature (Noble, Benjamin, Gebru) has been discussing what it means to encounter an ethical dilemma, and proposing possible solutions. Getting more diverse voices, and diversifying the kinds of thinking that goes into curriculums should be a foundational thought. Bears program, attaches undergrad with phd students at Columbia. Workshops every other friday, intro workshops to work as a space to troubleshoot and debut. Diversity in Computing series, speakers to talk about diversifying the industry from the inside and making the space more equitable.
Bednar: the program exists that there is no shortage of opportunities to study public policy, but they are fairly homogeneous and traditional. Professors of practice: what are we teaching about the messiness of public policy? For example, trying to recognize commercial interest in the privacy field — how do we take that on in the classroom setting? Bring in frontline voices to talk about how, when, and if we should be taking in business interests when designing interventions, and studying regulatory lag. Policy makers don't exist only to intervene when there is a problem - spice girl model of policy making, shouldn’t wait for polls from the people to actually work on issues.
Question: how do we integrate use cases in ethics?
Kelts: “How Humans Judge Machines”, used case studies from that. Used case studies from a training regimen at Google about self driving cars and face recognition app. Provide a schematized introduction about what the product does, and try to let them bring in the principle learned from the text, abstract it and apply it to the cases.
Question: what types of responsibilities fall on the data scientists vs. those assessing the models?
Akhtar: Chen from ParityAI, anti-racist AI startup, they are hired from corporations and act as third parties to assess biases in various automated models. Not everybody is trained to assess a model for a certain kind of bias, the responsibility should not be on one person. We need a multidisciplinary analysis of problems, and what that looks like in the real world. For example, what does that look like for credit resources and banks? It's important for people offering internships to show what this looks like in the real world, and where those opportunities may be.
Bednar: Skills labs, not training policy students to become senior data scientists. But they will be able to sit in rooms with those individuals and ask them questions about the design. You are developing a vocabulary, and a capacity for identifying context. Policy people should fit across multiple disciples.
Kelts: Value of assigning roles, but when are you going to intervene? By creating a sense of what’s going on in the room gives you context to make ethical judgments. Questions of privacy that were raised by early scholars of photography are very much the same of what we are asking today — those students had much more insight into what is also happening today.
Question: Will a national regulation be enough, or will we need a global one?
Bednar: the problems that we are seeking to address don’t sit well within national boundaries. In terms of privacy, we might see a global standard emerge, but its also important to see a federated approach. There will probably be more harmonization, but this could also just be a blip historically, and the real moonshot is eliminating regulatory lag.
Question: in addition to teaching policy theory, do you see a role for behavior science, and critical thinking?
Akhtar: the interest in responsible tech has exploded recently, and it correlates with a more public awareness of inequities within our society. If inequities that are built into our environment, if they remain with us and we don’t address them and the history that allowed them to rise, then how are we going to address those inequities in virtual spaces? Sometimes engineers will grab data from anywhere without understanding the source. How do we have a more holistic view of hard sciences, since they impact our daily lives?
Kelts: incorporate moral psychology into tech focused and policy focused ethics courses. This way people can start to contemplate how they want to counteract the biases in decision making and rationality.
Question: long term measures that should be executed now?
Bednar: navigating and negotiating private interests, more collaboration and overlap between business leaders, innovators, campus incubators, policy leaders, professional students, and how does that fit into regulatory realities. Big tech can be a red herring but we ignore the methods of how we can allow companies to grow in a way that is responsible.
Kelts: focus less on syntax or learning programming, and more on the design of the overall product. That is an opportunity for the ethicists and the policy people to enter into those classrooms to co-teach about the design of human societies.
Question: what would you want from a tech that is responsible?
Bednar: I want companies to talk to policy makers. Private actors dialoguing with governments should not be seen as negative, not only when it’s needed.
Akhtar: There is a slipperiness around topics of ethics and morality, we need to think broadly about who is in the room, who is coding. Also, we need to address issues of class, digital literacy and critical thinking should be built into education even before get to college, because that is not everybody’s life path. If we have a better understanding of inequities in our history early on in our lives, it will make for better candidates for whoever wants to enter the responsible tech space.
Question: take home message?
Kelts: tech should serve human needs, and not just human wants.
Akhtar: need to find a balance between a more democratic way of understanding what is responsible and what is ethical, and how do we support organizations that have a vested interest in making money but that can also deploy tech that is useful for humanity.
Bednar: policy makers and design innovators have more in common than we think, they are both problem solvers. We need to learn from people that are building things outside of government but still useful for the public.