Back in August, LinkedIn Learning released a new course titled “Data Ethics Fundamentals”. Produced by Texas-based software company Anaconda, Inc., the course is taught by software engineer Jose Mesa, Ph.D., a University of Michigan Engineering alumnus (Go Blue!).

I’ve taken online courses in data ethics before. They tend to focus on providing awareness to developers and engineers about ethics issues in tech and big data. As in, awareness that such issues exist. They tend to provide examples or case studies on topics like securing customer data from hackers or bias in AI models.

I expected this course to be much the same, but I decided to check it out anyway. What I found was a course that was more than topical, but rather provided methodologies for approaching ethical problems as you encounter them.

Definitions and paradigms

Philosophers are infamous for arguing semantics, with some debates over definitions spanning millennia. Mesa started the course with a straight-forward, working definition of ethics. This is a pragmatic move that makes the course approachable to non-philosophers (the intended audience).

Mesa introduced “ethics” as referring to the moral principles that govern a individual or organization’s behavior or how they conduct an activity.

Mesa proceeded to introduce four ethical theories that allow those working in big data to evaluate their own actions:

  • Deontology: “duty-based ethics”
  • Utilitarianism: “outcome based ethics”
  • Virtue ethics: “character based ethics”
  • Ethics of belief: “justification based ethics”

How to evaluate a dilemma

Mesa demonstrated how these ethical theories could be used to evaluate ethical dilemmas. In one example, Mesa proposed a hypothetical situation about a medical organization considering sharing its patients’ health information without consent for a study that could benefit public health. He then compared how a deontologist would respond differently from a utilitarian.

A deontologist, according to Mesa, would not share patient data without the patients’ consent as it would violate the rules. “Deontology focuses on the adherence of rules and duties,” he said. “Actions are considered ethical if they follow the rules, regardless of consequences.” To a deontologist, the potential for the data to help public health studies is irrelevant.

On the contrary, a utilitarian, Mesa explained, would be willing to share the data without patient approval if the potential benefits to society were greater than the issues concerning data privacy.

These different approaches can help those handling data in fields ranging from tech to business to medicine to education better evaluate their own actions.