Last year has seen some outcries over privacy breaches through NSA spying. We see more problems with pervasive monitoring and privacy through ad networks, Google tracking you, and soon possibly in your own home. This week we have seen that morality on the Internet goes beyond just privacy; there has been an outcry over the morality of Facebook manipulating news feeds for science. We need to look further than just privacy, and develop a new morality for the online space.
Facebook experiments with news feeds
Most users of Facebook have an enormous amount of friends, who post enormous amounts of updates, it is impossible to keep up to date on all of them. So Facebook implemented a filtering mechanism where you get almost all the updates of your closest friends, the ‘important’ updates from others, et cetera. Facebook also prioritizes these through some magic and orders them that way on your Facebook wall.
If you want to see everything in chronological order, you can select that, this will get you most updates from everyone. If you really want to see all updates from someone, you can always go to the wall of that person to see all their updates.
In early 2012 Facebook performed an experiment on 689,003 users, where they adjusted the filtering mechanism for some of these users. For one group they showed less positive updates and one group in which they showed less negative updates during one week in January 2012. The outcome of the study is that when Facebook exposes people to this kind of filtering, they follow that tendency and will post less positive, or less negative respectively (with only a very small effect).
This research was performed by Facebook scientists together with the University of California San Francisco (UCSF), and Cornell University. The outcome has been published in an article on the Proceedings of the National Academy of Sciences (PNAS). It has been online since July 2nd, but has only hit the news last week when it was published in print. The analysis below draws in part on the excellent coverage of The Atlantic.
Academic studies involving live test subjects should always be reviewed. In the US, universities have an Institutional Review Board (IRB) exactly for these kinds of reviews. The IRB of Cornell University has confirmed that it had no role in reviewing this experiment:
Because the research was conducted independently by Facebook and Professor Hancock had access only to results – and not to any data at any time – Cornell University’s Institutional Review Board concluded that he was not directly engaged in human research and that no review by the Cornell Human Research Protection Program was required.
The Facebook author of the study posted a public comment on Facebook describing how they collected the data and designed the experiment:
While we’ve always considered what research we do carefully, we (not just me, several other researchers at Facebook) have been working on improving our internal review practices. The experiment in question was run in early 2012, and we have come a long way since then. Those review practices will also incorporate what we’ve learned from the reaction to this paper.
To me this does not say that the experimental design was internally reviewed, and as a company they do not need to. The scientists associated to the universities simply treated this as a pre-existing dataset, and designed analyses of the data to be run by the people with access to the data, i.e. the Facebook author.
This way of dealing with the data could be seen as a sort of white-washing method where the ethically questionably acts have been performed by outsiders, and the analysis performed by the scientists.
Problems with the research
The problem that I have with this research is that it was performed on a randomly selected (large) set of users, without their informed consent. The American Psychological Association (APA) defines informed consent as:
When psychologists conduct research or provide assessment, therapy, counseling, or consulting services in person or via electronic transmission or other forms of communication, they obtain the informed consent of the individual or individuals using language that is reasonably understandable to that person or persons except when conducting such activities without consent is mandated by law or governmental regulation or as otherwise provided in this Ethics Code.
Facebook claims that this is part of their terms of service, which allows them to basically do anything with users’ data. However, they cannot reasonably claim that they ever asked for users’ permission in a language that ‘is reasonably understandable’.
It could very well be that how the data was handled and how the analysis was performed follows the universities’ guidelines, and that Facebook is allowed to do these kinds of things. However this does not mean that it is morally acceptable, similarly to the NSA spying. Especially academics should hold a higher moral standard.
These kind of studies and incidents show that ethical guidelines have to extend beyond the social sciences. Software and network engineers will increasingly deal with these kinds of issues, either directly by handling user data, or indirectly by creating systems that allow access to these kinds of data.
Most people are struggling with behaviour and data in the online space, as they did not grow up with it. I feel that we need to better understand what computers are doing. We also need to go one step further, and develop an intuition for the morality of designing and manipulating these systems. We have to figure out a way to teach this to our children, as they are the ones who are going to have to deal with this their entire lives.