Monday: Data Ethics
Wednesday: Data Ethics
Friday: Data Ethics Workshop.
- In Class as a domain group.
- Choose a hypothetical dataset associated with your domain (e.g., a dataset of faces and names for facial recognition). Try to be creative here. For instance, if your domain is politics, you might consider a dataset of voter information.
- Once you have decided on a dataset, apply the ethical considerations discussed on Monday and Wednesday for your dataset.
- Specifically apply the PFAT rubric.
- Answer the following questions:
- Privacy: Does the dataset contain any personally identifiable information (PII)? If so, how is it protected? If not, could it be used to identify individuals?
- Fairness: Could this data arbitrarily discriminate against any groups or be used to do so?
- Accountability: What are possible negative consequences related to the data, and how can you respond to them?
- Transparency: How would you explain the dataset to a user?
- Are there any values or assumptions embedded in the dataset? Could there be based on how it is collected, organized, or analyzed?
- End of Class:
- Present your findings to the class.
- Do not present your findings as a list of answers to the questions above. Instead, present your findings as a narrative that explains the dataset, the ethical considerations, and the implications of those considerations. Weaved into that discussion should be the answers to the questions above.
- Think of it as a short 3-minute presentation to a group of stakeholders.
Worshop Rubric
| Criteria | Points (0-5) |
|---|---|
| Does the group effectively discuss the fairness of the dataset? | |
| Does the group consider and foresee possible negative consequences of the dataset and respond to them? | |
| Does the group consider how to explain the dataset to a user? | |
| Does the group identify values or assumptions embedded in the dataset? | |
| Does the group narrate these considerations in a coherent and engaging manner? |