We will use an ethical framework to organize our creation of a gesture recognizer. First, we need to reflect on the problem we want to solve: how do we expect to use it? Who might it impact? What are possible benefits and harms? What values will direct our design? We will frame our project using the following problem statement.
Example problem statement
First, we need to identify the input data for the system. For our gesture recognition glove, we will need accelerometer data that reports the acceleration in the X, Y, and Z direction.
Second, we need to identify the output labels for the system. We can use accelerometer data to recognize positions and movements. Since this accelerometer will be placed on the hand, this means we can recognize hand positions and movements.
Let's consider a few possibilities for different output labels.
Recognizing sign language
Recognizing specific actions (e.g. like a smart watch or fitness tracker)
Recognizing gestures that might be used to interface with electronics
Ethics helps us analyze a scenario then systematically reason about different choices. We will envision possible uses for each project, identify key stakeholders, and identify some of the values help by stakeholders. Then, we will use this information to make informed decisions about what to design.
Possible Futures (envisioned use) cover the ways that a project might go into the world. This includes the scale of use and should point out potential benefits and harms of the work.
Stakeholders are any individual, group, or party that has an interest in an organization and the outcomes of its actions. This includes potential customers, their communities and ecosystems, corporations, their leaders and employees, investors and suppliers, and governments.
Primary Values describe, in broad and specific terms, the interests that various stakeholders have in the project. This can include considerations such as desired outcomes, preferred processes, adherence to moral or safety goals, profitability, and duty / responsibility.
Ethics protocol analysis of project ideas
Now that we've thought through possible futures, stakeholders and values let's choose one of these projects to work on.
First, let's consider our stakeholders, who we should include in our process. Many developers with great intentions (including the author!) have tried to build sign language translation gloves without appreciating that sign language is much more than a "hands" language - it involves the entire body. By including members of the Deaf community as key stakeholders in the creation of this project, we could avoid making this and similar mistakes. Unfortunately, I do not have access to these very important stakeholders at the moment. Therefore, I will set aside this idea for now and move on to the next one.
Second, let's consider our values. A wrist-based activity tracker can be a useful tool if it can predict activity accurately. However, detecting the difference between certain activities - is this vigorous wrist movement someone stirring a pot or taking a jog -might be very difficult. Accuracy is a high priority in this project and we might not be able to achieve it with the resources we have. Therefore, we will work on the third project idea.
Sign Language Gloves have a long, wraught history where engineers repeatedly design something that the Deaf community despises. The consequences of not including key stakeholders in the decisions about what and how to design systems include feelings mistrust, wasted time and resources, and perpetuating harm to already vulnerable communities.
"I was surprised and felt somehow betrayed because they obviously didn’t check with the Deaf community or even check with ASL program teachers to make sure that they are representing our language appropriately.” But after SignAloud received national and international media attention, Forshay teamed up with Kristi Winter and Emily Bender, from his department, to write a letter. They gathered input for the letter from the Deaf community and Deaf culture experts."
Our final problem statement is:
This algorithm will use accelerometer sensor data (x,y,z) to classify hand gestures like taps, rotations, and swipes.
If you chose to create a different project, document details about your design choices in the form below.