A dear friend of mine sent me a short email yesterday asking for some help and advice. He has been working on issues around human trafficking and has gotten invited to speak to Microsoft on the issue. He is a communication studies scholar with a background largely in media studies. So he asked for my help as the person that rubs shoulders with those computer science people more often.
So he asked: Given the current opportunities and challenges for researching and developing technological interventions in human trafficking it is now obvious that innovating solutions to human trafficking is inherently interdisciplinary and presents a great opportunity for computer scientists and social scientists to work together. From your experiences what would you say are the key challenges in working across social/computer science disciplines? Where do we complement each other? Where do we butt heads? How can we bridge the need for deep understanding of social dynamics via social science methods with the computational/engineering?
Of course, I can't give a short answer, but here are a few thoughts that I wrote back:
Social scientists do research often in service of theory, figuring out processes and practices - we are really bad at representing what we have learned in a language that people outside our specific fields would understand let alone find usable. Theories are just that - theories and they are very rarely infallible. They also tend to be overly broadly applicable to complex issues especially given the fact that the majority of research is done on midwestern college students. After all, in theory, theory and practice are the same... in practice they are not. How many times have you heard someone say: well "theoretically this is supposed to happen"... The reality of human life and human behavior that while some of what we do and how we react really does follow current theories, much of that applies to broad populations rather than being able to predict individual action. The trouble with the issue at hand though - human trafficking specifically - is that mistakes mean real broken lives and awful things.
Computer scientists are taught at the outset to "divide and conquer" - all problems are big and the only way to produce solutions is to start with small things and work from there. Trouble is, human systems don't work that way - divide and conquer results in reductionism of human behavior and can produce technology that has more difficult and negative side-effects than helpful positive effects (though at times we have huge wins, often in unexpected ways). CS needs quantification and human behavior, especially in the areas of things like human trafficking, is very difficult to quantify - it's an area where false-positives AND false-negatives can result in broken lives and awful things. HCI is a whole field trying to figure out how to deal with this - we no longer have humans and technologies, we now have socio-technical systems and that's really complicated.
There are some answers - big data in combination with ethnographic research, mixed-method and multi-method approaches to human behavior, technology probes and research through creating small disruptions in human lives to observe how socio-technical infrastructures function. All of these are current and ongoing methodological innovations in HCI and in tech industry research. We don't know if these will work, but it's really important to remember that big data is fallible and doesn't capture everything, that ethnography is so big and rich and detailed that it can make technology design impossible because there are just too many variables to consider, that technology probes can only happen for small self-selected samples and generalization from there is very difficult.
Human trafficking is something that happens at the edges, between the seams, intentionally camouflaged to remain unnoticed. Our current findings are often NOT generalizable precisely because this affects people that never make it into our research samples. There are so many challenges that worrying about disciplinary affiliation is stupid and expecting that one discipline (social science or computer science) can find solutions to the problems is simply ignorantly arrogant. So maybe it's just a call to action that to be successful, requires people to really put aside their disciplinary demagoguery and that's perhaps the hardest part...
No comments:
Post a Comment