Three ethical viewpoints applied to the Clinton email scandal

Last week, I gave an invited talk at the INFORMS Analytics Conference on ethics in data analytics. This is a follow-up to the HBR article from last year. In the talk, I discussed the need to have some frameworks to think about ethics. There are three main ethical philosophies that one can use to make ethical arguments.

***

The Hillary Clinton email scandal is a good place to apply these ethical theories. (See this Wiki page for a review of the scandal.)

Many journalists and people in the computer business are eager to ascertain if Clinton's use of a private server caused harm, as indicated by hacking. For instance, the headline of this New York Times article is "Security logs of Hillary Clinton's email server are said to show no evidence of hacking." This argument falls right in line with the "consequentialist" school of ethics. Followers of this school consider an action unethical if it causes bad consequences.

President Obama is not a consequentialist. We know this because he said, "Here’s what I know — Hillary Clinton was an outstanding secretary of state. She would never intentionally put America in any kind of jeopardy." His argument is that Clinton is a virtuous person and therefore her actions are ethical. This type of argument aligns with "virtue ethics," the second of three major schools of ethical philosophy. People who believe in virtue ethics do not judge individual actions.

By contrast, a consequentialist would not take intention into account. If you think you have developed a green product but it ended up making a thousand people sick, the consequentialist will still regard your action as unethical.

Others have said they lost trust in Clinton. This Quartz article talks about the violation of transparency and accountability. Trust, transparency, accountability: these are moral values that many of us treasure. Those who make this type of argument are engaging in "value ethics," which is the third major school of ethical thinking. Under this view, being ethical means one must follow moral norms of behavior.

Intention is clearly important to value ethics, as it is to virtue ethics. Consequence is of less import to either. The three philosophies overlap in places but disagree on some issues. When we discuss matters of ethics, we should have a clear picture of which general ethical approach we are using.

The field of data analytics is filled with ethical dilemmas. So I hope this overview is useful in sorting out your own thinking.

 

 

Comments

[Ken]

The legal system often uses a mix of schools. If someone attempts to kill someone and fails it is attempted murder and results in a lower penalty. Murder becomes manslaughter if it is not deliberate. Being an otherwise good person may reduce the sentence.

[zbicyclist]

A nit on philosophy #1 (cause harm). It's not just causing harm, but the likelihood of causing harm.

For example, if someone texts while driving and causes an accident, that's clear harm. But it's also wrong to text while driving, even if that particular time there was no accident, because of the likelihood of harm.

Apply this to Clinton. "There's no evidence of hacking" means we can't say there was harm, but we can say that this arrangement is likely to have raised the probability of harm given what seems to be a lackadaisical level of security. (In addition to the fact that hackers may just have left no trace.)

[Xiuyang]

This is really a good viewpoint of the data analytics. Now the data science is developing very fast and most of the ethical rules have not been established. Maybe we should really pay more attention to the ethical issues in the data analytics.

However, I think nowadays the building of an ethical system to supervise the data analytics is very difficult. In my mind, the most important problem for the data ethics is:

Talking about data ethics nowadays just likes talking about environment protection in 1950. We have not developed a complete data analytics system to utilized the data mostly. Which algorithm is best? How can we efficiently run a big data project? These problems are the most urgent ones which people want to solve. The professtional people have not enough time to build a data ethics system. There are a lot of utilization issues or problems for them to deal with. If these utilization problems cannot be fully solved, the outcome of the data analytics will be bad and then neither the data analytics nor the data ethics will exist. So now, maybe not many people will pay attention to the data ethical issues.

That is, I think, a contradiction or a balance problem. But most people will choose to focus on the productivity at first and then ethics.

Only some naive ideas, welcome to discussion. :)

Xiuyang