Josephine Wolff is Associate Professor of Cybersecurity Policy at Tufts University’s Fletcher School, a contributing opinion writer for the New York Times, author of You’ll See This Message When It Is Too Late: The Legal and Economic Aftermath of Cybersecurity Breaches (MIT Press), and she recently joined the editorial team of UC Press’s journal Global Perspectives as editor of the Technology and Global Change Section.
UC Press: Welcome to Global Perspectives!
JW: Thank you! It’s super exciting to be part of such an innovative and thoughtful interdisciplinary forum.
UC Press: How did you come to join Global Perspectives, and what was it that attracted you to the journal?
JW: I joined Global Perspectives as they were in the process of launching a new technology-focused section and I was primarily attracted to the idea of a journal that was not specifically grounded in a particular existing discipline. For all the emphasis on interdisciplinary research in academia over the course of the past decade or so, it’s still difficult to figure out–especially for scholars just starting their academic careers–where and how to publish work that doesn’t fit neatly into an existing field of research or fall squarely within a particular established department. This is something I’ve struggled with in my own career, as someone whose work spans a little bit of computer science, public policy, economics, and law, but doesn’t exactly fit the traditional mold of any of those disciplines. So I was excited to be part of something that was really designed to give people a place to publish rigorous research that touches on slightly unconventional combinations of academic fields.
UC Press: In your introductory essay, “How Is Technology Changing the World, and How Should the World Change Technology?,” you write about how an increased reliance on digital technologies and artificial intelligence can lead to our human biases getting embedded and codified within the technological systems we so rely on in the modern world. What are some examples of this risk?
JW: Perhaps the most publicized examples of this risk are the algorithmic decision-making systems that we increasingly find ourselves relying on, in both the public and private sectors, to help make decisions about how to translate languages, how to operate criminal justice systems, how to provide medical care, how to detect fraud, how to distribute social welfare benefits (to name just a few examples!). When we turn over even some of the responsibility for these decisions to artificial intelligence or machine learning algorithms we have to be really careful and conscious of the data sets that we use to train those algorithms. Because we know that if we train these systems on large corpuses of text, for instance, many of the biases underlying those text samples will then be embedded in those systems. Similarly, you can imagine how if we wanted to create algorithms that could help inform judicial decisions or credit scoring systems or employment decisions and we did so by training those algorithms on existing, human-made decisions in these areas, then all of the biases underlying those human decisions would inform the algorithmic decisions as well.
UC Press: What is the role of technology policy in addressing and managing these risks, and what is the state of our technology policy today?
JW: In general, I would argue the role of technology policy in trying to address and manage the risks presented by new or emerging technologies is to create incentives or requirements for the owners and operators of those technologies to put in place safeguards that their customers and users won’t demand, either because those users may not be aware of the risks and how best to remediate them or because they may not care about them. In cybersecurity (my own area of focus), that means requiring certain baseline security measures for critical infrastructure and sensitive data and networks–something we’re only just beginning to see policy-makers beginning to do. When it comes to biases embedded in AI, that will almost certainly mean requiring audits and transparency for the training data sets and outputs of high-risk AI systems, and it may also mean restricting the circumstances when AI can be used, especially in the public sector. That area of policy is at an even earlier stage of development than cybersecurity policy, so we really don’t know yet where it’s going or how long it will take to get there, but given how complicated modern machine learning algorithms are and how many companies are developing them, regulating them will probably be a slow and difficult process.
UC Press: Technology is certainly a huge driver of global change. How do you hope to inform our understanding of technology through your section of Global Perspectives? What kind of papers would you like to see published, and do you have any words of advice to those who are considering submitting manuscripts?
JW: My hope is that this section will grapple directly with the technical elements and design of emerging technologies and how those specific design elements inform the impacts of new technologies on society. To me, the most interesting social science research focused on technology is the work that looks at how the technical architecture of a system for transmitting information or bioengineering or generating energy (or anything else) directly informs that system’s impacts and which elements of that design are susceptible to various kinds of change or influence. I’m hoping the work published in this section will answer questions about technologies like: Are there design decisions we can make, at the technical, regulatory, or social levels, that might help shift or improve some of the ways that technology is changing the world, and if so, how feasible are those decisions, and who would need to make them?
UC Press: Thanks again, and best wishes for your section of the journal!
JW: Thank you!
Global Perspectives is an online-only, peer-reviewed, transdisciplinary journal seeking to advance social science research and debates in a globalizing world, specifically in terms of concepts, theories, methodologies, and evidence bases. Work published in the journal is enriched by invited perspectives that enhance its global and interdisciplinary implications.
online.ucpress.edu/gp