Are ethical considerations for the data scientist within scope? Is {ethics} on topic?

Question

Are all ethical considerations for the data scientist already subsumed by a blend of other professional disciplines and Information Security (INFOSEC)?

I can see that many ethical dilemmas will be related to INFOSEC or corporate espionage and data privacy. While ethical questions could be pushed off to other sites, is there a place for those questions here? Does the exploitation of big data create for us any professional responsibility?

I would argue that we do have professional responsibilities, but many of those have to do with our concepts of privacy, ownership especially information ownership, and public welfare. And since these concepts are highly dependent upon our cultural upbringing, I wanted to invite discussion on this topic to expose those things that myself and others may be taking for granted because of uncritical cultural biases.

If we do have our own professional code of ethics (even if in the making), where do these unique responsibilities and concerns originate from?

If our concerns are not terribly unique, then is there something remarkable in the collection of concerns that occur regularly for data scientists that warrants expounding or codifying?

A question of scope

If there are ethical considerations that are unique or are more than just the sum of other disciplines, then I would propose that questions about ethics are on-topic for this site.

If questions about {ethics} is on-topic, does it need its own tag? And/or should we adjust the site's help page?

However, if the prevailing opinion is that there is nothing really specific to the ethics of data science that warrants separate treatment - then we can determine such posts to be off-topic.

Answer 1

The ethics of dealing with data, particularly data that is tied to specific, identifiable users, is an interesting and important topic. But I would argue it's outside the scope of the site.

I think the site should be focused on questions which professional data scientists are especially qualified to answer. Data scientists might be more qualified to pose ethical questions relating to data (i.e., thinking of ways data could be used that might not occur to others), but I don't feel that this makes them any more qualified to answer such questions. Moreover, I think that any such question could be presented in such a way that the potential ethical implications were clear to non-data scientists. Given this, I think those questions ought to be asked on an ethics-focused site, not a data-science focused one.

It's not just a matter of keeping the focus of this site from getting too broad, but also of getting the best answers to those ethical questions. With questions of ethics I think it's especially important to get a broader perspective, not limited to those working in a particular field. Consider how the average government intelligence agent's perspective on the ethics of government surveillance might differ from that of a professional ethicist, or from the broader community, and whether it would really be better for the government agent to only be posing such questions to their peers in the agency.

Also, I think that keeping the site focused on data science theory and practice will keep it more in line with what people would likely expect the site to be about, and more consistent with the level of focus on other sites across the network. A question like "should physicists accept defense department funding, even if they don't approve of their government's military commitments overseas" would probably be deemed off-topic on physics.stackexchange, and I would expect people to have a similar expectation here.

Answer 2

@TimGoodman makes a strong argument; I'm going to try to present a counter-argument.

Tim raises the question: "Are data scientists especially qualified to answer ethical questions relating to data?" I think we agree that the answer is no, not necessarily; but is that really an appropriate metric for whether such questions are on-topic?

Professional data scientists are more likely to have experience and training relevant to certain ethical questions surrounding data. Because it's a somewhat interdisciplinary field, many of these questions are more appropriate for a different site on the network, e.g.:

• Ethical questions about storing data --> security.SE or DBA.SE
• Ethical questions about sourcing data --> opendata.SE

Still, we can reasonably expect that there are ethical questions intimately related to that intersection of fields where data scientists live. We shouldn't a priori dismiss such questions as outside the scope of a community explicitly intended for those who have a personal or professional interest in just such a discussion, on the basis that not every data scientist is also an ethicist.

I would argue that the best answers to such questions are more likely to be informed by professional experience in the field of data science than by any other experience.

Every profession that involves interaction with customers or the public demands that its members spend some time during their careers thinking about ethics. Professional organizations have ethical standards and perform ethical inquiries; certification and licensure exams include ethical questions.

On the other hand, how many people with careers unrelated to the field of data science are obligated to spend time "exercising their data science muscle?" It may be that the best answers to these questions will come from someone who is both an ethicist and a data scientist, but we're probably more likely to find that person here than on a general ethics site.

In fact, there is no ethics site in the SE network - ethics questions currently belong to philosophy.SE. I don't think it's debatable that the proportion of users who have at least some understanding of both data science and ethics and a desire to discuss these questions is much higher here than it will ever be on philosophy.SE. Perhaps at such time as there is a professional ethics site in the network, we could revisit this question, but for now I think it's in our best interest to treat ethics questions as on-topic, but migrate them to closely related sites like opendata.SE or security.SE on a case-by-case basis.