Before this class I thought…
- Designing ethically is important but is a difficult and overwhelming task as you attempt to consider all the things that could go wrong with a technology.
- I thought conducting studies was complex and challenging.
- Encouraging users to change their behavior through technology must be done in a manipulative way.
About the experience…
Initially, I wasn’t crazy about the ethical discussions and felt we might be spending too much class time on them, but as they continued, I grew to look forward to them and found them extremely valuable. I believe this class truly prioritized ethical design, which I appreciated. My experience with ethics in CS has been in my problem-set classes. In these classes, ethics has felt more like a distracting joke. However, I believe ethics makes sense in classes such as CS147 or this class because ethics in CS is all about applying ethics when designing systems. In classes like CS103, we’re simply learning the fundamentals of implementing algorithms and using computer systems, and ethical questions presented in these classes never felt useful. However, in this class, our discussions we’re beneficial. I found that I was able to apply some ideas we had in this class about how we can monitor pervasive organizations that implement persuasive designs in another class of mine where I had to consider how we could help big tech create tech that does less harm and has less unintended effects. The idea I was able to incorporate was that there should be a UN Committee to regulate tech companies and asses and critique major technologies. This allows representatives from different countries where these technologies are used to have a forum to express their concerns with emerging tech.
I took this class while taking another class called Stanford Climate Ventures (SCV). SCV is a GSB-centered class where I got to work with 3 Ph.D. students, two MBA students, and a JD student on creating a business plan based on how to commercialize a lab-scale technology. The technology can turn plastic into high-value chemicals. During this class, we conducted 70 interviews and worked on building the business model from the ground up. During the process, we decided to conduct a survey to get user feedback on how they would feel about chemicals made from plastics and other things. While I was creating this survey, I realized… it’s an assumptions test! It’s an assumptions test because we were trying to understand if our assumptions about customers and those who would engage with our business were correct.
For instance, these were two assumptions we had:
- People won’t think chemicals made from plastic waste are yucky since most chemicals currently come from petroleum.
- People don’t read the ingredients much in general, but they read ingredients for personal care products more so than cleaning products.
We tested these assumptions by creating a google form survey and distributing it. There are example questions from the survey in the image below.
Overall the project was very smooth. I was blessed with a great team who were all motivated to produce high-quality work, and I feel like the prototype is finished, and we did the best we possibly could. I believe the biggest challenge was navigating how to give feedback in a constructive way and learning how to approach our previous team member, who we felt wasn’t delivering meaningful contributions to the team. I loved that Cristina encouraged us to give feedback constantly to each other in class and check in. I found her feedback/conflict resolution lecture extremely helpful. Besides using her techniques when giving constructive feedback to our previous team member, I also used them during my RA interview (and I got the job (: ! ). In the RA interview, we were given prompts about how we could help facilitate conflict resolution between roommates. I made sure to emphasize that individuals often may not realize certain behaviors are bothersome to others, and when delivered in a thoughtful way, feedback can be liberating and help people understand how they can improve. I found this lecture very helpful and a life skill I believe they should teach in other classes.
As someone who is chronically late and has a lengthy self-care routine, measuring me helped me come to terms with the reality of how much time I spend doing self-care and how much time it takes me to get ready. It encouraged me to give people an accurate estimate of how long I will take to get ready or when I will arrive at an event, whereas before, I would disillusion myself because I would be ashamed it would take me long to get ready, so I wouldn’t know how long on average it would take me to get ready because I never wanted to face the truth. I will definitely be doing a quarterly “Measuring Me” activity from now on to check in with my self.
I found the mapping activities early on a bit confusing (fishtail and connectivity circle). I wish the connectivity circle was explained in more detail. I felt quite confused about where to start for this activity and what the nodes around the circle of the map should be. However, this has the ability to be a really powerful tool for people if instructions for how to create it is a bit more clear. I also thought that it was a bit overwhelming being presented with the fishtail, connectivity circle, and bathtub draining visualizations all at once. It might be a good idea to go in-depth on one visualization, so people understand it rather than trying to teach them all.
Ethical considerations…
Our app takes several measures to ensure the user’s privacy, such as not selling data and allowing users to delete data. In the privacy readings, we learned that the premise of privacy behind the “I have nothing to hide argument” is that privacy is about hiding a wrong. However, privacy isn’t just about hiding wrongs. It’s about concealing information you are uncomfortable with others knowing, and it’s about protecting yourself from others using information against you. This is why our app collected the minimum amount of required data because unnecessary data collection and use are not helpful for our users. In the future, it will be important that as AI technology evolves, we only use AI technology that analyzes the bare minimum amount of data to meet the app’s needs. For instance, currently, the AI is examining the photos of individuals to determine if they’re on or off task. This means the AI doesn’t also need the ability to interpret text data. Instead, it should simply be optimized at reading people. Although it may be helpful for the AI to read text data because it could interpret signs or text in the images, it’s an unnecessary feature that doesn’t align with the purpose of the AI, which is to determine whether the user is on task or off task.
To change behavior, we nudge users to meditate when they want to refocus after returning from a break by making it the default option. It’s an acceptable nudge because they can skip it or choose a different refocusing activity, and meditating is scientifically proven to be beneficial, so choosing to meditate to refocus is the rational choice. The intervention could become manipulative if the skip button is barely noticeable. As the app evolves, it will be important to continue to ensure the “skip” button is available and accessible otherwise, the design will become manipulative because it would deceive users into thinking there’s no other option besides forced meditation.
Now I think…
After this class, I realize designing ethically isn’t just about damage control, it’s about proactively integrating ethics into each step of the design as a value to optimize for in the design process, and this makes it much easier to create an ethical solution. I also found that it’s easier to think ethically about designing a solution the more I practiced.
I also feel more confident conducting informational studies and intervention studies, and I believe they’re a valuable tools when designing a solution. They help you gather general information and understand if your interventions are effective before spending lots of time and money to construct a full prototype.
Finally, I realized designing for behavior change doesn’t need to be slimy and manipulative. This is just how big tech companies often implement behavior change solutions. By using participatory design and ethical design guidelines, one can create an app that can change its user’s behavior without it being manipulative.
Next Time…
The next time I am given a chance to design a solution, I will make sure to identify my assumptions and conduct assumption testing if needed, and I will be careful about implementing my solution in an ethical way and encourage my team members to have an ethical discussion about our design at each step in the design process because the more frequently we have the conversation, the easier designing an ethical solution becomes.
