When Facebook was revealed to have been experimenting on the emotional state of 700,000 of its users back in June 2014, many were outraged that the company had violated ethical guidelines and “harmed” its users.
The fallout of that “emotional contagion study” haunts Facebook’s reputation among ethicists and researchers. But it has also compelled the company to clean up its ethical act, and inspired the introduction of a newly developed internal ethics review process this week.
In an article in the Washington and Lee Law Review, Facebook’s public policy research manager Molly Jackman and research management lead Lauri Kanerva discuss the challenges of devising an ethical review process at Facebook and detail the new process research must follow.
Ethical research, the piece claims, is now integrated into the workflows of its developers and researchers. “The model best suited to protect people and promote ethical research is one that fits the unique context in which the research takes place,” write Jackman and Kanerva. This means that asking ethical questions should become “part of researchers’ normal workflows” rather than an extra step or additional burden on their work.
Facebook’s set of suicide prevention and support tools, which rolled out globally this week, are a case in point. The tools were developed in collaboration with researchers at the University of Washington and directly serve the kinds of vulnerable and sensitive users the review process is designed to benefit.
This approach goes some way towards meeting calls from researchers and others who believe that ethical considerations must be more than a tickbox in the development cycle of a product or service – or an add-on after it is released. These considerations should be infused throughout the lifecycle of research, from conception to analysis to dissemination of results.
Facebook’s response to the emotional contagion outrage has been rather more sophisticated than the social network’s early days; Zuckerberg told upset users to “calm down” and “breathe” after the controversial introduction of the NewsFeed in 2006. Ten years later, the new ethics review model demonstrates a company that is maturing and coming to terms with the gravity of its role as the largest – and most impactful – social networking site on the planet, with 1 billion users every day.
But how does Facebook translate broad ethical values of respect, diversity, beneficence and justice, for example, into its decision making for research around specific audiences, projects or products? From Facebook’s new process, it is not entirely clear.
Tailoring processes too closely to a specific company risks serving the company better than the subjects in need of protection.
Tellingly, Facebook’s descriptions of procedure and process offer little insight into the values and ideals that drive its decision-making. Instead, the authors offer vague, hollow and at times conflicting statements such as noting how its reviewers “consider how the research will improve our society, our community, and Facebook”.
This seemingly innocuous statement raises more ethical questions than it answers. What does Facebook think an “improved” society looks like? Who or what constitutes “our community?” What values inform their ideas of a better society?
Facebook sidesteps this completely by saying that ethical oversight necessarily involves subjectivity and a degree of discretion on the part of reviewers – yet simply noting that subjectivity is unavoidable does not negate the fact that explicit discussion of ethical values is important.
One of the reasons we have standardized ethics codes and discussions is so researchers, administrators and ethicists alike can think through, communicate, and justify their decisions – however subjective.
Laura Stark, who has written about institutional review boards (IRBs, the standard oversight mechanism for ethics review for academic and publicly-funded research) explains that researchers consistently use subjective reasoning when considering the ethics of a research proposal, often asking questions like “what would my grandmother think if she were asked these questions?” or “how would my cousin respond to this study?”
But, Stark points out, this kind of ethical deliberation could vary wildly depending on the composition of the board; one whose members are uniformly affluent and white is going to arrive at a different assessment than one that is racially and socioeconomically diverse.
Nothing about Facebook’s review process gives users and would-be research subjects insight into whose ethical subjectivity and standards they are being subjected to at any given time.
In the absence of any meaningful insight into Facebook’s ethical commitments, the authors note that their review process allows them to “seamlessly” call on representatives or groups outside of Facebook for consultation as needed – though it’s never made clear why or how this additional external input is triggered.
It’s also not clear how Facebook selects external groups to contribute to their reviews. Internal research involving potentially vulnerable LGBT subjects points vaguely towards consulting “prominent groups representing LGBT people”.
But which LGBT groups? Not all groups or organisations serve the interests of the L, G, B, and T equally well. Mainstream LGBT advocacy groups such as the Human Rights Campaign are often the most visible (HRC has worked with Facebook in the past), but has been criticised for often prioritising the interests of cisgender and largely white gay men while failing to account for the interests of transgender folks and people of colour.
So, if you’re the most vulnerable of vulnerable research subjects, it’s still not clear when and how your interests might be promoted or protected. Nor is ity clear how users can advocate for their interests or gain insight into research that might target them.
Ultimately, Facebook has offered an ethics review process that innovates on process but tells us little about the ethical values informing their product development. What we have gained is insight into process designed to allow Facebook and its researchers to pursue their research agendas while avoiding public controversy on the scale of the emotional contagion episode.
But without a substantive, transparent debate about ethical standards and far more detail about Facebook’s own values, this isn’t really a process about ethics – it’s just PR.