Teaching Responsible Computing Playbook

Teaching Responsible Computing Playbook

Effective Talking Points for Different Audiences

Authors: Subbu Vincent

In the context of Ethical Computing, because it is a nascent field and an emerging interdisciplinary area, talking points emerge over concerns, questions, worries, conflation, divergent meanings, and lack of a common vocabulary, assumptions, perceptions about the disconnect between philosophical theory and application into computing practice, jargon and so forth. For instance, discussion about “fairness” in moral theory is far less anchored around the technical/definitional boundaries implied in training datasets. But the latter is a helpful tool in computing to build fairness as a procedural constraint for ethical-AI.

People want to talk about these things to ensure they get on the same page with each other in a conversation about ethics and computer science, be it designing a project, experiment, course, study session, community event, competition, etc.

This section will describe some common talking points, how to anticipate them, address them and use them to engage both faculty and students. We will first list terms that have emerged in the responsible computing journeys of several universities, contexts the points arise in, and what concerns and tensions underpin them. Then we will outline some ideal scenarios where the talking points get addressed or receive ventilation or demystification. We will also outline a checklist to follow and some key questions to cover.

Developing effective talking points and strategies for applied ethics on computing is necessary because most of us (whether students or faculty) already have a personal idea of what we think the word/words mean. Still, we worry that no two people are going to agree on it, or that ethics is subjective, personal, overly values-driven, etc. Equally worrisome is when two people talk about ethics, but frame “right or wrong” around very different moral foundations without realizing it, since the training or exposure required to verbalize is missing. For instance, one person may feel a decision is wrong from a human rights perspective, while another feels it is right on fairness. Unarticulated concerns and objections may be present in our minds that ethics does not fit technical contexts where the disciplinary boundaries of right and wrong are otherwise self-evident (Algorithms optimized for speed, memory, reducing technical debt, etc.).

Talking points help with reframing ethics to address concerns people may have but not articulate, and yet will prevent us from fuller engagement. Below are some examples of talking points:

  1. What is ethics? How does the computing faculty broach this? How do students?
  2. Isn’t ethics for students, whether computing or otherwise, just about “not cheating”? How do ethics get into a technical area like computing?
  3. What is not ethics? Is ethics the same as religion? Is ethics a feeling about right and wrong?
  4. Can applying ethics to computer science or computing, in general, be done given computing is an objective technical discipline? Or not?
  5. Isn’t ethics too soft? Computing undergrad students innately feel technology is amoral and can be used for good or bad, a hard science. Why bring ethics into it at all?
  6. What are some examples of activities that are called ethical computing?
  7. Does ethics integration compromise computer science or computing training in some way? Does it dilute rigor in learning?
  8. Let’s take a concept like fairness. Computing folks already know to implement “fairness” (e.g. round in algorithms for resource allocation, proportion, etc. Why make a big deal out of this?)
  9. How can we even measure ethics learning in the computing context? If two students get top-grades in a Computing class, and one does poorly on ethics aspects, does that mean that student is a poorer computing student?


Key Questions:

  • How might one frame ethical thinking? Keep the following as guidelines for framing ethical thinking (they reduce the sense of threat and uncertainty and fit the learning aspirations of students):
    • As a skill
    • As part of professional development
    • Self-learning makes a person-professional whole
    • Gaining a fuller understanding of the impact and implications of product work, release and harm mitigation in a heavily interconnected world.
  • Is there a plan in place to handle objections? Some examples of push backs are: “ethics is not technical”, or that “If I can’t have objective measurable learning outcomes, I don’t know how to do it”, etc. The key is to prepare in advance to handle objections.
    • One starting point is to say “There are always multiple legal or technically valid ways to implement a user experience feature or algorithm”, and ethics help decide which is the right approach for a particular situation.
    • Identify handy computing-ethics cases. Cases constructed as stories usually showcase why there are always multiple legally or technically valid designs (from things as simple as a data structure holding gender, to a case where a synthetic text generator might spit out hate text at the user). Software engineering ethics cases are abundant.
    • Have some material around the difference between “dense philosophical theory and reasoning” and “applied ethics”. Integrating ethics into computing classes and activities are examples of “applied ethics” where the moral principles and lenses (developed in different strands of philosophy) are applied to specific domain level problems.
    • Another starting point would be to say that ethical thinking is a kind of brain muscle development that happens through ongoing exposure and regular exercising, like anything else. It is part of self-development.
  • What is the plan to set expectations initially for a new computing class that integrates ethics in it? At the start of a full semester class in computing (say cybersecurity or data science) where ethics is being integrated, can the instructor plan a few minutes initially in-class mentioning that ethical thinking is being introduced? At this stage, the instructor can give a few general tech-ethics reading resources for orientation, and a “what is ethics” reading piece. This will help inform students (if they forgot about this aspect of the course, rather than discover that ethics is part of class 3 or class 4.)
  • How does one account for worries that the students will feel their faculty is a computing expert and not a philosopher, but ethics is a philosophy, and why they must pay attention? If the computing instructor is integrating ethics for the first time in a course, can the instructor identify a speaker from your computing colleagues or applied ethics or philosophy with a bridge-applied background, who can lead the conversation about “what is ethics and what it is not” instead of the instructor? Having a guest speaker for one or two specific sessions is helpful.
  • Can the students be engaged in ethics and computing only through courses/classes? Less-formal venues may provide open conversations about ethics that start with “What does ethics mean to me?” where everyone can voice in, without feeling like any answer is “wrong” or “right”. However, it is useful to have a structured response to that question. Students’ associations on campus (ACM, IEEE, etc.) may already have monthly or quarterly meetings with good outreach. If so, a student leader may be willing to propose whether evening student community events can be held on computing ethics cases. (1-hour events, discussion, with case walkthrough and pizza)


Checklist

☐ Create a framework for ethical thinking

☐ Create a plan to handle objections

☐ Plan for setting expectations for a class where ethics is being integrated

☐ Make sure students hear from faculty who are experienced in talking about ethics

☐ Create avenues beyond classes to engage students in ethics


Examples

Santa Clara University

At Santa Clara University (SCU), we decided to do student community events in the evenings to discuss a specific case study that raises ethical questions on privacy, or data bias, etc.

Checklist walk-through: Below we show how some of the checklist items in the previous section of this document could be used at SCU. (The checklist did not exist when the assignment was designed and given. This walk-through is meant for illustrative purposes.)

  • Create a framework for ethical thinking
    • In each event, we use the applied ethics orientation literature here to talk about what ethics is and not. We find that this sets expectations clearly that ethics is not formulaic, considerations include multiple moral lenses, stakeholders, power, and decision making. Our framework is posted here.
  • Create a plan to handle objections
    • The Markkula center has a suite of software engineering ethics cases. You may also use the points in second Key Questions above to state objections upfront and handle them.
  • Create avenues beyond classes to engage students in ethics
    • We had the fortune of being able to work with the help of a student engineering fellow at the applied ethics center to liaison with the computing/engineering students community and organize. In this 7-minute SCU RCS video, we've outlined the content flow for such an event, examples of cases, and a checklist for planning. Briefly:
      • How to decide a format for a student ethics event: Students often prefer hands-on, relevant work relating to real, ethical challenges in product design or software engineering. We decided to do community events because they are off-curriculum, and students can discuss the case together without the trappings of a classroom environment. Case study discussion events based on recent events offer a chance to expose them to ethics and recent developments/controversies in one effort. They complement curricular ethics integration efforts nicely.
      • How did we decide topics?: Many stories break into the tech-news sphere every month. Topics range from cybersecurity, privacy, bad actors gaming data science, etc. For instance, two case-discussions we picked were the Goodbye Fears Monster, and Open AI's GPT-2 release. The latter was a major story in 2019. Here is a brief checklist (we've published this in our video as well.)
        • Decide an ethics framework to help walk through the case
        • Identify a case and write up no more than a two-pager
        • Plan the agenda: expectation, an introduction to ethics, time to discuss the case in breakout groups, for students to weigh after discussion
        • Plan for food and time of day.
        • Avoid > one hour, or 65 minutes, tops
        • IMPORTANT: Block 5 minutes for students to fill the feedback form before they leave.
        • Questions for event feedback form:
          • Determine what the students felt they were learning
          • Which parts they found most illuminating and why


Resources


Related Pages


Authors and Contributors

Subbu Vincent

Subbu Vincent (author)