In the book Value Sensitive Design, author Batya Friedman defines value sensitive design as a theoretical approach to design rooted in human values. She explains the need for designing technical systems in a way that brings human values to the forefront, while also honoring existing and well-functioning technical efforts.
In my Interaction Design Studio, my project group is designing a virtual assistant for Blue Apron (BA). On a high-level, our design goal has been to provide a more personalized, hands-free cooking experience. Chef, BA’s virtual assistant, guides users through meal cooking, makes smart meal suggestions based on user data, and connects to smart kitchen appliances/third party apps.
Direct and Indirect Stakeholders
Below is a map of human and nonhuman stakeholders who might be affected, directly or indirectly, by Blue Apron’s VA/Blue Apron’s system.
Benefits, Harms, and Values
The following are listed by key stakeholder groups:
- values that might stand in conflict with these benefits and harms
Blue Apron’s largest customer base
Benefits: A virtual assistant (VA) provides a hands-free, personalized cooking experience. Our team is making an assumption that this customer base uses Blue Apron because they lead busy lives and need an easy, quick, and efficient way to cook.
Harms: Data. Personal data is collected and used so that the VA can provide a personalized experience. This connects to values of privacy, control, and autonomy. Users might want a more personalized experience, but need to give Blue Apron access to more of their personal data. Users can be tracked based on habits, location, dietary preferences. This also means that a user’s data now lives on a server somewhere in the world.
Values: Eating healthy food, ease, efficiency, personalization, connection, privacy, control over one’s own data
Benefits: Partners might benefit from a personalized cooking experience that’s efficient and easy. VA might also facilitate a connection between partners.
Harms: Personal data being used, tracked, and saved.
Values: Eating healthy food, easy, efficiency, personalization, connection, privacy, control over one’s own data
Google NHM, iPhone X, Smart Appliances, Third-Party Apps
Benefits: Communication with other devices make each device smarter and more efficient.
Harms: Data collection and potential gaps for companies creating APIs for different devices to communicate with one another. Battery life of devices might also be affected. Technical glitches may mean more dev or customer support resources.
Values: Reliability, efficiency, autonomy
Benefits: If more users adopt BA because of its personalized, efficient experience, chefs can make a living. Promote sustainable and high-quality recipes.
Harms: User preferences dictate recipes. Limits creativity if users can make substitutions or users might choose similar recipes continually.
Values: Sustainability, quality recipes, earning a living, creative control
Benefits: Sell crops to Blue Apron business. Possibly earn a sustainable living if one of BA’s values is promoting an ethical/sustainable food system.
Harms: User preferences dictate farming. Monocrop culture if users order the same meals or continue getting substitutions from grocery stores.
Values: Sustainability, earning a living, healthy farmland, control over crops planted
Benefits: Organic products create more fertile, healthy, long-lasting soil.
Harms: If there is a shift to mono-crops because of user preferences, it’ll be bad for the land, soil, and ecology of the area.
Values: Diverse ecological environment, sustainable farming practices
Benefits: To shift towards more ethical, sustainable practices.
Harms: Radical change might cause the system to break down (people lose jobs, protests, becomes a divisive political issue). Change BA is promoting might create a more unequal, inaccessible food system (ethical food system for those who can afford it).
Values: Sustainability of both food and people who support the industry
Employees who prepare/deliver meal kits
Benefits: Earn a living
Harms: Schedule dictated by users, control
Grocery stores/grocery store employees/food suppliers
Benefits: Continue earning a living/getting business. If BA offers suggestions about picking things up from local grocery stores.
Harms: More users adopting BA meal kit service, which means less business for grocery stores. Missing the internet/data trend, where data gathering is seemingly a very important thing in determining the success of a business.
Values: Earning a living, control
Key Value Tensions
Personalization vs. Privacy: To provide a more personalized experience, Blue Apron requires user data. Users have to give up this data, allowing it to be stored on external services and used to track their behaviors, habits, and location. That’s pretty invasive, standing in tension with user needs for privacy.
Personalization vs. Accessibility: Personalization requires data collection. Data collection requires technology. Blue Apron’s service, smartphones, and smart kitchen appliances are not cheap. Blue Apron’s value of accessibility stands in tension with the value of providing a more personalized (more costly) experience.
User Control vs. Autonomy: Does providing a more personalized experience mean that the AI can make more decisions for users? Or should users still retain a certain degree of choice and control? What is that degree?
User Control vs. Collective Control: Providing a more personalized experience means that users’ wants might dictate company decisions. How does this affect chefs who create meal kit recipes or farmers who grow meal kit crops? Will there be a shift to growing food and creating recipes based solely on user preferences? What impact will this have on ecology, the food system, or social structures? How will this impact the livelihoods of employees?
Sustainability vs. More Technology/Current System(?): Technological solutions are not always the most sustainable ones. Further, sustainability for whom? For people who can afford the BA service? How does this affect BA’s vision for a more sustainable food system — if access is unequal and uneven? Will this end up creating an even more unequal, unsustainable food system?
Lots of Assumptions
Value sensitive design calls for a design process that is rooted in a comprehensive set of investigations — conceptual, empirical, and technical investigations.
Right now, our groups are navigating within and creating work based on highly conceptual investigations. It would be great to do the empirical work needed to confront, edit, and refine the many assumptions we’re making.