OUR LIST OF POINTS:
1) the normalization of using metadata (in the us vs. elsewhere)
2) how big data or metadata becomes a way of "knowing" for policymakers/researchers
3) and how that process is in itself problematic (nontransparent, possibly manipulated and driven by interests in the region)
From a technical and very procedural perspective:
* She claims the data is anonymized. There are many different claims that are bunched into that statement (and let's discuss together if I am reading too much into this):
*a) That both her IRB and the source had some sort of consent from the participants of the study
*b) That she indeed could anonymize the dataset: we know from research that data anonymization is impossible, data anon in network and location data is a myth
*c) She has some sort of informed consent from the callers in participating in this study, i.e., the whole country of Yemen concented and were well informed of the potential uses of this data, including potential military and intelligence use.
*d) She seems to then do ananlysis that is specifically targeted at identifying communities and even profiling individuals. This contradicts the suggestion that she is looking for general patterns rather than working on profiling individuals or small groups.
*e) She doesn't give source or provide any suggestion as to whether her collaboration with the military means that this dataset and her analysis is available to the military which is part of the current war. (that last bit needs to be said more elegantly). Did the IRB accept this and how exactly did they evaluate that providing data to the military is ethically in the interest of these people?
*
And, at the end of the day, suggesting that this is a infringement upon privacy misses all the points that Arun said.
*
*
COMMENTS FROM Arun/Paula/Miriyam/Joris:
Arun:
+++++++++++++++++++++++++++++++++
Some quick notes based on a rough read - all pretty obvious stuff to everyone here I imagine.
The “problem” that big data is meant to solve: places that are “hotspots” require “us” (i.e. the West) to have knowledge about them but they are also “inaccessible” to Western knowledge because it’s hard to do “social science” there. The hidden assumption here is that Yemenis cannot represent themselves but need to be represented by Western academics, journalists, NGO workers, etc. (Meanwhile we ignore Yemenis who come to the US to testify about the impact of drone strikes, etc.)
The desire for knowledge of Yemenis: "You know where he lives. You know who he calls and who his friends are. You know if he’s Shia or Sunni, depending on what holidays he makes calls. You know if he’s rich or poor depending on how much phone credit he uses.” This is “knowledge" that is inseparable from the “policy” uses to which it will be put (i.e. imperial violence to suppress political movements or to “reform” practices such as Khat consumption). This “knowledge" is funded by, framed by, and used by empire, and only meaningful in those terms. I guess one outcome from the military’s point of view (who apparently funded this work) is that data on the patterns of cellphone calls after a drone strike will be fed into social network analysis to indicate additional potential targets.
What is Big Data actually doing in this study except just confirming the anthropological assumptions that the researchers begin with? Isn’t Big Data just putting a quasi-empirical gloss on essentialist assumptions about Yemeni culture?
Obviously, this is fundamentally an issue of academic complicity. Just as there have been movements by psychologists and anthropologists to say that those disciplines cannot participate in the War on Terror by assisting in torture or the military’s “Human Terrain Systems” programme, so too the same battles will now need to be fought in the fields of computer science and data science.
Miriyam:
+++++++++++++++++++++++++++++++++
Thanks for taking the lead on this guys, I remember our anger at some
of the utterly shameless self-promotion on airl list around something
similar. What really strikes me here is also the completely obvious
indifference about the actual really devastating miseries in what is
mainly US/KSA crafted conflict turning violent after a mass popular
uprising was highjacked. Yemen is probably the bitterest example of
the ugly liaison between military intervention and local power
struggle. The country, its people its institutions...all going down
the drain (read Helen lackners pieces). Not a single sentence is
reserved for that in this example you shared. Heartless emotionless
careerist gibberish which will probably be defended as Scientific
objectivity'.
Maybe we should organise a second infrastructures of empire and focus
on this area of imperial research. Plenty of examples as joris says.
An event that will also expose this would be my preference. Think to
get Omar involved, he has looooots of examples of academic complicity,
geography, Palestine, bds
Paula:
+++++++++++++++++++++++++++++++++
I would just suggest two points in terms of development and geopolitics. 1) Cold War precedent for failed and contested theories of propaganda/surveillance exported wholesale to 3rd world beginning in post-coup Iran. So this region has long been US technology theorists laboratory of failed experiments. To me this is Daniel Lerner and company redux. 2) The MIT poverty lab model is based on finding technical "non-political" solutions to development. This is a perfect example among many others in their portfolio that brackets politics and ultimately any power relations for "win-win" solutions.
Important to point out systemic nature of this kind of framing not an aberration.There is lots of work of this vein around the 3rd world post conflict landscape.
Good luck on the piece and we need to publicly shame this kind of scholarship and institutions that support/make possible like MIT and many more.
Joris:
+++++++++++++++++++++++++++++++++
Related, there is a lot of talk going around on humanitarian data and tech innovation:
It ranges from the openly awful:
http://www.timesofisrael.com/how-big-data-in-the-backroads-of-africa-will-help-feed-the-hungry/
to the cleverly framed (data innovation should smarten up otherwise it will backfire)
http://cis-india.org/papers/ebola-a-big-data-disaster
A good many of the folks interested in data and humanitarian development are on the responsible data forum list.
http://lists.theengineroom.org/lists/info/responsible_data