Students Engage with Website Transparency
Professor Essick and Professor Edenberg lead the class.
Human-Computer Interaction (HCI) took center stage in the Responsible CS Team’s third session in Ray Essick’s Advanced Programming course. The class thought through a broad range of ethical questions related to the ways in which users engage with websites and apps, raising issues of transparency, consent, manipulation, as well as accessibility for individuals with disabilities. By reflecting carefully and critically on the ethical implications and tradeoffs implicit in digital interface design, students learned and practiced skills of responsible programming that will help inform their careers.
In advance of class, students completed a homework assignment in which they chose a website or app and considered a range of HCI-centered themes, including how easy or hard it is to find information; whether choices are structured and presented in manipulative ways; for whom the interfaces may be intuitive and for which groups they may pose challenges of access and navigation; whether it is easy or hard to tell if data are being collected; and whether data collection is opt-in or opt-out. Students also reflected on what is the most important thing to fix with the interface from an ethical point of view.
Values and Tradeoffs
Students came prepared to discuss HCIs related to shopping, food delivery, travel, academia, and other topics. In a lively exchange co-led by Professor Essick and Professor Elizabeth Edenberg, Senior Ethicist at Ethics Lab, the class raised incisive points and asked difficult questions about the values and tradeoffs inherent in HCI. When it came to a popular video streaming service, students observed that it was hard to tell how one’s data may be used. Conversation about whether privacy policies and terms of service are easy to understand led to questions about the limits of legitimate consent. Professor Essick asked how many students carefully read terms of service to activate the “Agree” button, rather than scrolling to the end, and few said that they do. One student pointed to a subtle interaction between updating terms of service and lock-in effects: The student noted a difference between using an app for the first time and choosing whether to accept new terms of service for a site where one has developed an extensive social network and presence over time. Someone else drew a comparison between visual HCIs and auditory HCIs. Whereas with visual interfaces users can see a menu of options and site maps, voice-enabled HCIs “just have options presented back to the user,” so the device “is running the interaction whereas with a visual interface a user is running the interaction.”
Professor Edenberg encouraged the class to think more deeply about the perspectives of diverse sets of users who may be differently situated from programmers who design HCIs, particularly individuals who have visual or hearing impairments. As she noted, “By listening to people you can design much better systems from the get-go.” Along those lines, one student gave an example of a clothing website that she thought scored well on accessibility, with an icon of a person displaying a menu of options, including reading what is on the site, minimizing blinking, and other features.
Broadening the range of factors computer scientists think about in designing HCIs to include ethical considerations about ease of use, comprehensibility, degree of meaningful choice, power dynamics, and accessibility for those with disabilities helped students expand their sense of their professional obligations as computer scientists.
This project is funded by a Mozilla Responsible Computer Science Challenge grant and is part of the university’s Initiative on Tech & Society, which is training the next generation of leaders to understand technology’s complex impact on society and develop innovative solutions at the intersection of ethics, policy and governance.