6 Ethics
Stacy Reardon; Rachael Samberg; and Timothy Vollmer
An overview of ethical considerations in TDM
How can digital humanities researchers incorporate an ethical framework into text data mining? Privacy protections, which are primarily legal mechanisms, will only get us so far. As Todd Suomela and colleagues wrote in Applying an Ethics of Care to Internet Research: Gamergate and Digital Humanities,[1]
“Humanists are not used to thinking ethically about research subjects because we mostly deal with either subjects who are dead or subjects that are public figures like authors, politicians or other humanists, whose roles and activities are open to scrutiny and debate. The real or potential harms inflicted by research methods on these subjects are often intangible and hard to measure.”
We will explore two key ethics questions by surfacing theory through case studies and scholarly literature.
First, we learned in the chapter on privacy that when “public” data is being collected and republished for TDM purposes, it isn’t necessarily protected by privacy statutes. So, in the absence of privacy law requirements, to what degree of care should we treat TDM data?
Second, if the current regulatory framework for research involving human subjects is set up to protect privacy, how do we know when to impose an ethical framework? And when and how do we balance the application of a framework with truth-seeking, the public interest, and free expression?
Public, but sensitive
So what do we do with data that is not technically private, but we feel might be sensitive?
As we saw in the Gamergate research mentioned above, the bringing together of public social media messages can serve as a signal boost for hate messages and misinformation. The mere act of compiling these Tweets makes these messages and the targets of the messages more easily discoverable. Yet in many cases, there is nothing private in this data, at least as far as privacy is defined by state and federal laws. But aggregating these hateful messages can also boost the signal concerning views about which an author later has changed their mind. So we have to ask, what is our responsibility as researchers?
Second, should we somehow account for the fact that creators of data might not understand what is protected by privacy law, but might think that data they make available online should essentially still be treated as private? After all, it has been repeatedly demonstrated that the average user’s expectations of “privacy” don’t necessarily match with what the law says. So do researchers bear an obligation to adhere to social norms or ethics to protect it?
Mismatched expectations are a particular problem when data crosses international boundaries. Many big data research initiatives are international, and protections vary substantially depending on which national data protection regulation applies. Research subjects may believe that the regulations of their home country protect their personal data, when