Ethical Critique of Human-Computer Interfaces

In the second sessions in Professor Essick’s Advance Programing course, the Responsible CS Team tasked students with ethically vetting human-computer interfaces by identifying how specific design choices can affect software users for better or worse in ethically salient ways. These engagements extend the previous sessions’s insights regarding how computer science and its applications are infused with value. And, importantly, these sessions took place before students learned about the computer science of human-computer interface design itself, attuning them to the ethical considerations of upcoming course content. Along with Professor Ray Essick, Ethics Lab’s Senior Ethicist Professor Elizabeth Edenberg and Postdoctoral Fellow Mark Hanin led the sessions which were also notable for being virtual rather than in-person, in light of the COVID-19 pandemic.

To prepare for the session students completed a homework assignment in which they scrutinized the human-computer interface of a website or application of their choice in light of factors such as its intent, accessibility, and data collection settings. As they engaged with the interface, students documented their ethical critiques of specific design features using screenshots, text annotations, and other visual markups. Professors Edenberg and Essick, and Postdoctoral Fellow Hanin, reviewed the assignments in advance of the session, preparing a powerpoint structured around key ethical highlights that emerged from the analyses. The virtual session itself was a dynamic blend of lecture and discussion that made use of screen sharing, white boards, powerpoint, and chat to examine several ethically relevant features of human-computer interfaces that students identified in their investigations.

Identifying Ethical Concerns

One major cluster of concerns students had about the interfaces they examined centered on user control, default settings, and manipulation. Many students observed that hard to find and difficult to navigate settings menus often resulted in users opting for the software’s default settings. Professor Essick reminded students of their responsibility as designers to choose default settings wisely with the user’s wellbeing in mind, while Professor Edenberg and Postdoctoral Fellow Hanin cautioned how defaults that are difficult to find or adjust can lull users into a false sense of security and threaten their autonomy. Of course, manipulative practices are not unique to software design. Citing JC Penny’s coupon campaign as an example of the anchoring effect and the placement of cereal boxes on grocery store shelves as an instance of target advertising, Professor Essick encouraged students to reflect more deeply about whether and how manipulative practices in the digital realm differ from real world varieties. One student argued that digital manipulation may be more pernicious: unlike product placement in stores or coupons appearing in circulars, advertising on screens is constant, ubiquitous, and highly personalized, ultimately having a greater impact on a person’s preference, choice, and sense of well-being.

Students also raised concerns about data tracking. Part of their worry was that, for many sites and applications, it was unclear what data was being tracked and how it was being used. Sometimes the lack of clarity was due to privacy policies that were exhaustive or difficult to understand. As Professor Edenberg and Postdoctoral Fellow Hanin observed, the opacity here is an instance of the transparency paradox—the tension between presenting complex information in a way that is both accurate and comprehensible.

One student saw this phenomenon as a good place to locate an ethical principle: if a company cannot easily explain what it is doing with someone’s data, that may be a good enough reason against such use.

Some students expressed less concern about their site or application’s data tracking noting a lack of obvious bad effects as well as some improvements to functionality. For instance, with data tracking enabled on the application for McDonald’s, one student observed more accurate information about nearby locations which he found useful. In light of such tradeoffs between functionality and privacy involved, Professor Edenberg urged students as both users and designers to be guided by consideration of what data tracking is genuinely needed to provide users with value, while Postdoctoral Fellow Hanin reminded students of the potential for the unexpected. Even if a single application’s data tracking does not pose any risks now, that data may be combined with other data in ways that pose risks to individuals in the future.

The accessibility of human-computer interfaces was another major topic of concern. Many students noted how poor choices in font size or coloring may needlessly prevent older or visually impaired individuals from using an application. One student’s experiences growing up abroad posed deeper challenges to the assumptions underlying human-computer interface design. Growing up where online banking and cashless transactions were not secure and where mail services were less reliable, he reported that Amazon’s services were effectively inaccessible. Postdoctoral Fellow Hanin encouraged students to appreciate the deeper lesson from these cases of inaccessibility.

All of us may take [basic aspects of daily experience] for granted because of the particular community we live in” but, to make applications widely accessible, designers should seek to use their “moral imagination” to access the perspectives of people who face radically different circumstances.

By scrutinizing human-computer interfaces with the user’s interests and perspective in mind, these sessions were an engaging way for students to appreciate the ethical dimensions of interface design choices especially as they relate to autonomy, manipulation, data tracking, and accessibility. Recognizing some of the ethically salient considerations at stake in the design process is a key step in the responsible CS curriculum. Reflecting on the ethical complexities of interface design, Professor Essick concluded the session by reminding students of the importance of this ethically responsible first step.

We’re not asking you to think about these questions because we believe that you’ll have all of the answers… The important bit is that you see the possibility of more than one answer and say ‘that’s something to ask, that’s something to discuss.