The last two weeks you read about how power shapes and determines the types of data we collect in ways that are not always visible or beneficial for society. This week, we will explore some example datasets to try and understand how power and bias are embedded in the data we collect and use, as well as in determining what we don’t collect. We will also explore how we can use cultural data to challenge and critique power structures.

Group Activity

Once again, each group should first create a Google Doc and share it with the class in our Slack channel. Please name your Google Doc “Critical Cultural Data Explorations - Group [Number]”.

Your goal for this assignment is to work together to answer the following prompts:

Cultural Data Exploration

  • Find a dataset that contains data about historic cultural objects or practices (historic meaning created prior to the 1980s). This could be a dataset that contains information about a specific cultural object (like medieval manuscripts) or a dataset that contains information about a specific cultural practice (like historic art collecting), etc.
  • Find a dataset that contains contemporary cultural data (so data from post-1980s). This could be a dataset that contains information about contemporary cultural objects (like movie genres) or practices (box office grosses or ratings for movies).

You can find more than one example for each of these, and you can also try to find cultural data that spans both time periods, or use different types of cultural data (like music, film, literature, etc.). You should include a link to the dataset and a screenshot of the dataset in your Google Doc.

For each dataset, try to answer the following questions in the Google Doc:

  • How do these two datasets differ in how they represent cultural objects or practices? How much can we understand of how this data was created or collected? Do the two datasets differ in how they were created or collected?
  • How does this dataset reflect power structures? How does it reflect the biases of those who collect and use the data? Does this data challenge or critique power structures in any way? How does it relate to the concept of matrix of domination from the Data Feminism reading?

Finally, the group should consider what is missing from these datasets, and why that data is missing. Is that data missing because it is difficult to collect, because it is not valued, or because it is actively suppressed? What would it look to collect that data and how might it change the dataset? Should that data be collected?

Cultural Data in AI Exploration

Using an AI Chatbot (whether that’s ChatGPT or some other chatbot), ask the chatbot to generate a description or story about your contemporary and historic cultural object or practice. What sorts of languages does the chatbot use to describe your objects or projects? How does the chatbot’s text reflect the biases of the data it was trained on? How does it reflect power structures? Does it challenge or critique power structures in any way? Does the chatbot generate the same types of descriptions for the historic and contemporary cultural objects or practices? And does it generate the same results for everyone in the group? Finally, how does the chatbot consider what is missing from the data? Does it generate descriptions of what is missing?

Post screenshots of the chatbot’s responses in your Google Doc. We will discuss these responses as a class.