7 Ethics and Social Impact

7.1 Privacy and Surveillance Discussion: Reidentification and Facial Recognition

Data about people inevitably brings up questions about privacy and surveillance. While the word “surveillance” may conjure images of security cameras and hidden microphones, today all kinds of organizations are collecting data on us through digital devices of all kinds–some in our pockets or living rooms.

Even for those who “have nothing to hide”, this sort of surveillance is concerning.

As data scientists, analysts, and decision-makers, you will be faced with the opportunity and challenge of handling data about people. You may decide what data to collect, how to store it, how to share it, and how to use it. “Acting justly” with this data requires thinking about, among other things, the right and responsibility of privacy.

Pick one of the following two topics (Reidentification or Face Recognition) for this Discussion. Then:

  1. Pick an article or video to engage with.
  2. Write and post your observations and questions. Things you might mention:
    • a one-sentence summary of the article for your future self
    • a quote and your thoughts about it
    • something you found surprising or cool
    • something you disagree with
    • something you didn’t understand
    • a question this raises for you
  3. Read what your peers wrote about the other topic.

7.1.1 Reidentification

Suppose your organization is asked to share data it has collected about individual people. You might think that removing the names is enough to protect privacy. Is it?

Read this overview: https://georgetownlawtechreview.org/re-identification-of-anonymized-data/GLTR-04-2017/. Focus on sections 1 (Introduction), IV (Re-identification), and V (Conclusion).

If you want to go deeper (or watch a video):

  • Watch the video on this page for an example of the kind of concerns that come up when trying to anonymize data.
  • Bonus: read Sweeny's original paper for context

7.1.2 Face Recognition

Face recognition is useful to people: auto-tagging on social media and photo apps, unlocking your phone, convenience in boarding planes or paying for stuff. And it’s useful to business and government: tracking customer behavior in stores, finding criminals, identifying people critical of the government, etc… and many people think that some of these uses cross an ethical line. There is also evidence that many facial recognition systems perform more poorly for some groups of people, hence increasing the risk of harm to them. Others think that the potential for harm or abuse makes facial recognition technology something that should not be deployed even if it were perfectly accurate.

Face recognition technology has recently gained more visibility as companies like Facebook have announced they’re discontinuing some forms of face recognition.

Our question:

How might facial recognition technology be useful? how might it be harmful?

7.1.2.0.1 Articles or Videos

Going deeper:

  • The 2021 film Coded Bias is feature-length discussion of facial recognition and related issues.
    • The IndieLens discussion guide has a lot of summary information, background, and references.
  • See the UMich ESC Project Green Light site for some other articles.

7.2 Background

7.2.1 Current Issues