Mining the Military-Academic-Industrial Complex in a Poetic-Serious Fashion
- Data Mining, Collaboration, and Institutional Infrastructure for Tranforming Research
- 2 positions: Data Mining Consultant and Social Science Data Consultant
- CFP: Special Issue of the Journal of Educational Data Mining
- new article: "Trending: The Promises and the Challenges of Big Social Data"
- HASTAC seeking Postdoctoral Scholar for NSF EAGER Social Network Data Grant
Sadly I was not able to attend the HASTAC conference, but I wanted to post some information about a current project of mine...
As we now hear by some commentators that the "worst" of the so-called financial "crisis" "might" be over, we have to acknowledge the difficulty of squaring these remarks with the realities of our colleagues in the university. Friends are not having their teaching contracts renewed and graduate students in the humanities are having difficulty in finding TA positions for the fall semester. I recently overheard one administrator in a more "technical" department suggesting that the lack of available TAships in the humanities is due to their inability to support students on "research" grants. He found no problems with the fact that his own department was able to admit five new graduate students this year, while a similar-sized humanities department can barely afford to admit one student. Survival of the "fittest", indeed.
In these times, then, perhaps we ought to be focusing our critical lenses on the institution of the University itself, to better understand its role in the relations of capital. Thankfully, many are doing this, and the recent and current protest actions at schools in the US and around the world are evidence of one, very much needed, type of response. My own response has been to develop MAICgregator (http://maicgregator.org), a Firefox extension that aggregates information about colleges and universities embedded in the military-academic-industrial (MAIC) complex. It searches government funding databases, private news sources, private press releases, and public information about trustees to try and produce a radical cartography of the modern university via the replacement or overlay of this information on academic websites. MAICgregator is available for download right now:
http://maicgregator.org/download . If you want to see what MAICgregator does to a website without downloading it, you can look at some screenshots: http://maicgregator.org/docs/screenshots . This is its first public release, so expect that things might not work properly.
I have written an extensive statement about MAICgregator that tries to contextualize it within discourses of net.art, the military-academic-industrial complex, "data mining", and activist artistic practices. I include this statement at the end of the e-mail, but it is also available on the MAICgregator website:
I welcome any feedback or discussion that this might provoke; if you want to e-mail the project authors directly, please e-mail info --at-- maicgregator ---dot--- org.
Mining the Military-AcademicIndustrial Complex in a Poetic-Serious Fashion
MAICgregator is a Firefox extension (http://www.artwarez.org/femext/index.html) that aggregates information about colleges and universities embedded in the military-academic-industrial (MAIC) complex. It searches government funding databases, private news sources, private press releases, and public information about trustees to try and produce a radical cartography of the modern university via the replacement or overlay of this information on academic websites. This is a necessary activity in light of the contemporary financial ?crisis?.
net.art that will not be version numbered
Firefox extensions (putting aside (sadly) for the moment the regrettable continued use of male language) or add-ons are today presenting a relatively low-barrier entry into the development of web- and browser-based artistic projects. While we do not want to discount the level of programming knowledge necessary to build them, they are still based on a libre platform and can be developed within a rather large community of programmers, programmers whose own work is, by default, available for inspection and study. (What we mean here is that all of the source code for an extension is contained within the extension itself, making it easy to learn from the work of others.) This is in marked contrast to previous (and contemporaneous (and future)) strands of net.art that might have valorized the use of Director, Shockwave, Flash, or Java, the first three being expensive, proprietary, and closed platforms, and the last being an open programming language, but one where the actual source code is often difficult to get to if it is not provided directly by the artist.
Recently there has been a flurry of add-on development (http://artzilla.org/) both poetic and serious. We would be remiss to ignore the work of Michael Mandiberg (http://www.mandiberg.com/) who was involved with two important early extensions, Oil Standard (http://transition.turbulence.org/Works/oilstandard/) , a project that would replace dollar amounts on pages to their equivalents in barrels of oil, and Real Costs (http://therealcosts.com/) , an add-on that shows carbon for alternative forms of transportation on major airline websites. Similar in this vein is the recent Add Art (http://add-art.org/) plugin by Steve Lambert (http://visitsteve.com/) that replaces advertisements with curated net.art shows; Track-me-not (http://mrl.nyu.edu/~dhowe/TrackMeNot/) , that floods Google with fake searches to descrease the efficacy of data mining; turkopticon (http://turkopticon.differenceengines.com/) that allows people to see and respond to shady Amazon Mechanical Turk employers; and China Channel
(http://chinachannel.hk/) that lets people surf the web behind the Great Firewall of China. Besides these there are add-ons for ?censoring? text on a page using black blocks (http://www.gleuch.com/projects/ctrl-f-d) as well as returning us to the heyday of mid-1990s web design (http://timemachine.6x.to/) . This brief list of add-ons shows the extent to which artists are re-purposing the web browser for radical artistic purposes using technology that is (potentially/oftentimes) much easier to work with than in earlier net.art times.
While there might have been other ways to develop and release MAICgregator, it seemed most appropriate to develop it as a Firefox extension, especially given the ability in many places to (at least temporarily) install these plugins on public machines at colleges, universities, libraries, or internet cafés. The ease by which extensions can be installed?and the ability a programmer has to modify pages at will once the extension is installed?makes them an ideal vector for the propagation of radical or alternative perspectives to those that are fixed on the web page itself. This is especially the case within the modern university, where schools carefully control what types of information make it to their front page or internal portals, or where students consume their news in computer-generated chunks via Google News, absent marginalized or alternative voices. Add-ons provide one way to break open this lock on web-based media, combining disparate sources together in a montage that is at once both serious and poetic.
Obviously a plug-in like MAICgregator is not going to immediately alter materially the construction of the military-academic-industrial-complex. Nevertheless, in just the short time we have been developing it, we have come to a much better understanding of the links between these major actors, as well as some of the more strange players in the mix (http://www.marketwire.com
/press-release/UrologicalcareCom-958964.html) . Part of our intent is, yes, to provide carefully selected ?facts? about the relationships between universities, the military, and the corporate world. But just as equally our goal is to perform an alternative, to show how so-called ?news? sources can be recombined in new ways to create novel connections. This is a performance that additionally re-opens the consideration of exactly what the ?web? is, given its continued atrophy into staid configurations of a media-controlled semiotics. And it is finally a performance of what we might call a ?poetic austerity?: the use of whatever means are available to us?absent the possibility of funding through traditional sources, given their decrease in this time of focusing on the ?essentials??to respond to power on our own terms. While it might be feeble, while it might seem utopian (http://brechtforum.org
/who-will-build-ark-utopian-imperative-age-catastrophe) , it is certainly necessary nevertheless to do what we can, with what we have, against that which oppresses, by creating that which we want.
The Military-Academic-Industrial-Complex (MAIC)
MAICgregator exists as one attempt?of many?to counter the hegemony of the present-day University. It is well-known that former US President Dwight Eisenhower, in his 1961 farewell address to the nation, wanted to speak not only of the military-industrial complex (http://en.wikipedia.org
/wiki/Military_industrial_complex) that became the renowned phrase that it is, but rather of the military-academic-industrial complex (The University in Chains: Confronting the Military-Industrial-Academic Complex
(2007), Henry Giroux, pp 13-15), what we are here calling the MAIC. Former US Senator J. William Fulbright (http://www.countercurrents.org/us-turse290404.htm) spoke publicly of the same thing in 1968. As you visit US college and university campuses today, you easily see the extent of the military-academic-industrial complex. Companies and academia are big business, and while the ?end? of the cold war took away the spotlight from the relationship between academic and the military, ?defense? monies easily find their way into the university and out again to defense contractors. Corporations fund endowed professorships, schools outsource fundamental operations such as their bookstore to corporations like Barnes and Noble, and universities offer advertising space on brand-new plasma screens installed in said bookstores. Ads for everything from spring break vacations in Mexico to jobs at Lockheed Martin plaster the walls of today?s campuses.
This is all the more heinous now given the interrelationships between universities, corporations, and capital. The precipitous fall in university endowments is linked not only to the contemporaneous use of business models in the governance of universities, but also on the transplantation of corporate fund managers to highly-paid positions in the university hierarchy. This is an invasive transplantation, as fund managers more accustomed to the risk profiles of hedge funds are ill-equipped to managing much more conservative portfolios such as those of a university, an institution that is predicated on continued existence without the possibility of being ?sold? or ?broken up? in any type of ?bankruptcy? proceedings. Indeed, it has been reported that Harvard University?s endowment was, at one point recently, leveraged 105 percent (http://www.nytimes.com/2009/02/21/business/economy/21harvard.html) ?meaning that it had invested more than it actually had on hand. While this may be a common tactic of those who come from a corporate finance background, it becomes downright distasteful in the context of an institution such as a university. Yet the University jumped on the bandwagon of hedge funds, real estate speculation, and investment in private equity, enticed by the thought of big returns. However, as soon as the investments went down, the endowments tanked as well.
While the day-to-day functioning of the University in the United States is vested in the oLces of the President, Provost, and various Vice-Presidents, the overall strategic direction is governed by a Board of Trustees or, in the case of public universities, Regents. Private colleges and universities are actually chartered as non-profit corporations and, as such, legally require these Boards in order to exist. These Boards function quite similarly to their counterparts in the corporate world, the Board of Directors: trustees have final say on all tenure decisions, they set fund-raising goals, decide on capital projects, and help set the direction of endowment investment. Thus, they are also implicated in the horrible decline in endowment monies experienced by many schools in the last year. However, their activities and deliberations are done almost entirely in secret: while there is a token movement towards transparency in the convening of public ?forums? during regular trustee meetings, most proceedings are done behind closed doors. Such lack of transparency has been one of the most prominent issues raised by the recent protesters at NYU (http://takebacknyu.com/2009/02/19/nyu-occupied/) and the New School (http://newschoolinexileblog.blogspot.com/2008/12/our-demandsfor-negotiat...) .
Indeed, the links between schools, corporations, governments, military activities around the world become rather frightening once you start putting it all together?which is of course why there is a lack of transparency in the first place. For example, Cornell University has received gifts from former Citigroup CEO and chairman Sanford I. Weill totalling in the hundreds of millions of dollars (http://www.news.cornell.edu/stories/April09/WeillGift.html) for its medical school in New York City. Citigroup, through its subsidiary Citibank, has additionally been involved in providing loans to a rebel group in Democratic Republic of Congo (DRC) associated with the brutal civil war (http://www.business-humanrights.org/Documents/Vuyelwa-Kuuya-on-UN-Expert...) in which millions have been killed that is additionally funded in part through the sale of coltan (http://mason.gmu.edu/%7Ejmantz/Improvisational%20Economies.pdf) , a mineral that eventually is transformed for use in high-performance tantalum capacitors like those used in modern electronic equipment. There are reasons why some people do not want these sorts of links to be made?
Nevertheless, these links do need to be made, and are being made everyday by the powerful through the practice of data mining. Data mining as a term is a remarkable bit of rhetorical slippage or slight of hand. In the juxtaposition of two terms we see the elision of disparate meanings and the movement of concepts from one word to the other. ?Mining? used to refer primarily to the material, the digging into the earth in order to extract something of value, something that was hidden on the surface but became seen only through the hard labor of others?immigrants or the poor?in order to be sold as raw commodities used in the production of further commodities in the chain. Iron, gold, diamonds, copper, tin, aluminum, coltan?these are things that are mined. They can be held in your hand or in the back of an enormous truck. Mining creates land disputes as ?rights? are now bought and sold for the contents of that which cannot be seen, but which can be sensed through various forms of technologies that can ?penetrate? the earth. Mining is the creation of gashes in the earth in order to further our appetite for other items in which the mined material does not ?appear? at all. Mining is still a vital component of the world economy and can be especially harmful and contentious, as we have seen with coltan and as threatens to happen with lithium in Bolivia (http://news.bbc.co.uk/2/hi/business/7707847.stm) . Data, on the other hand, is perceived to be the most immaterial. It is the thing that can travel ?instantaneously? from one place to another, that has no physical analog, that does not obey physical laws. Yet data is materiality at its most fundamental: it always already exists as magnetic bits on a platter, or the movement of electrical or optic pulses down a wire or fiber; it is subject to the same physical laws as everything else in the universe. Data?s meaning ?other than in its form as abstraction?is always imposed from the outside; the bytes that make up a text file are meaningless without a lookup table that says the number ?68? represents the letter ?D? or the number ?100? represents the letter ?d? (and yes, case is important or ?sensitive?). Can these mappings be ?owned?? Can data become a commodity?
Of course these are questions that have obvious answers today, and it is partially due to the rhetorical power of a term such as ?data mining?. The phrase itself thus brings the legal power of the owning of mining ?rights? to immaterial ?data?, creating a mongrel that at the same time diminishes and displaces the horrors of the continued physical mining that must take place in order to feed the machines needed for ?data mining? itself. What a concept! The list of situations in which corporations, the military, and the government use data mining is enormous and growing constantly. To take a couple of recent examples:
* The Defense Advanced Research Projects Agency (DARPA) created the Total Information Awareness (http://epic.org/privacy/profiling/tia/tiasystemdescription.pdf) (TIA) program in 2002 that, under an ?Information Awareness OLce?, would ?imagine, develop, apply, integrate, demonstrate and transition information technologies, components and prototype, closed-loop, information systems that will counter asymmetric threats by achieving total information awareness?. While many found their logo (http://en.wikipedia.org/wiki/File:IAO-logo.png) to be beyond creepy, it was the danger of this mass surveillance system that brought the program to a halt after extensive proceedings by members of the United States Congress.
* Acxiom (http://www.acxiom.com) is one of the largest and most powerful of the companies that market data products (http://www.acxiom.com/dataproducts) to corporations that help them ?segment? ?customers? on a variety of axes. One of the more interesting ?products? they offer is a report on the underbanked (http://www.acxiom.com/77348 /UnderBanked_Indicator) which, in their words, ?helps marketers find potentially profitable underserved consumers who lack formal banking relationships and represent an untapped pool for checking, savings, fee-based, prepaid or starter credit services.? This is done through mining various types of public databases and connecting that data with information from the credit bureaus to find those that are not in the latter.
These data mining endeavors function on the premise of perfect, or near-perfect data: that the data they produce, use, and sell gives an ?accurate? picture of the situation and cannot be gamed or interpreted differently. Yet, as anyone who has taken a statistics course knows, ?garbage in equals garbage out?. Data mining rests on an enormous, shifting mound of assumptions that can be endlessly debated and tweaked. And the data that forms the ?inputs? to these models can additionally be manipulated, through simple changes to the name (as those subject to misplacement on terrorist ?watch lists? know all-too-well) or flooding of data aggregation sources with ?wrong? data.
Data mining, or the more common term in the academic community, machine learning (where we should obviously put the word ?learning? in quotes), is an incredibly hot topic these days, especially given the proliferation of social networks and the enormous amount of data to ?mine?. Of course there are the obvious privacy ?concerns? that are brought up in papers (http://portal.acm.org/citation.cfm?id=1281192.1281195) but with little consideration to other alternatives (such as not collecting the data in the first place, or radically reconsidering social network research to not be faced with such privacy concerns). Academic researchers from across the world receive funding to ?mine? mobile phone call records (http://portal.acm.org/citation.cfm?id=1244002.1244212) , ?analyze? how people revisit web pages (http://portal.acm.org/citation.cfm?id=1357054.1357241) , or characterize ?types? within particular configurations of networks (http://portal.acm.org /citation.cfm?id=1117454.1117457) . This sort of research is replete with actors from industry, including Microsoft Research (http://portal.acm.org/citation.cfm?id=1357054.1357241) , HP Labs (http://portal.acm.org /citation.cfm?id=1150402.1150423) , AT&T Labs (http://portal.acm.org /citation.cfm?id=1242572.1242579) , and Dow Chemical (http://portal.acm.org /citation.cfm?id=1081870.1081970) , companies that are all able to present at mainstream academic conferences put on by one of the two main professional organizations in the computer science field, the Association for Computing Machinery (http://www.acm.org) (ACM).
Nevertheless, we can consider an alternative form of data mining, one we might want to call alter-data mining. Taking into account all of the caveats that we have already mentioned, as well as the conceptual issues with data mining in and of itself, we might be able to turn data mining techniques on the powerful themselves, using the results to being to form one alternative mapping of the situation, while in the process commenting on the role of data mining in society. This is the main conceptual foundation of the MAICgregator project: that perhaps we might be able to aggregate some of this data and through direct and poetic presentations of it, turn it into actionable information.
Performing the ?re-? on ?data mining?
Of course the aggregation of disparate sources of ?data? or ?information? or
?signs? is not a new technique in the arts. It is certainly older than the Dadaist?s photomontages, but it is interesting to turn to them, and especially the work of Hannah Höch, as it provides a historical parallel to our re-appropriation of military and commercial techniques:
"Actually, we borrowed the idea from a trick of the official photographers of the Prussian army regiments. They used to have elaborate oleolithographed mounts, representing a group of uniformed men with a barracks or a landscape in the background, but with the faces cut out; in these mounts, the photographers then inserted photographic portraits of the faces of their customers, generally colouring them later by hand. But the aesthetic purpose, if any, of this very primitive kind of photomontage was to idealize reality, whereas the Dada photomonteur set out to give to something entirely unreal all the appearances of something real that had actually been photographed." (Hanna Höch, Interview with Edouard Roditi (1959), Dada, Phaidon Press, 2006, edited by Rudolf Kuenzli, p. 232)
This reapplication of techniques for radically different purposes is a trope in radical and avant-garde artistic practice, yet it continues to have an effect with each new form of media. From re-purposing the spectacle to reapplying the nomadic war machine, performing the ?re-? or the ?trans-? has the potential to create new configurations that upset present balances of power.
In recent networked and online works there has been much interest in taking these techniques of data aggregation and re-appropriating and applying them to the powerful actors. Designed as a response to the Total Information Awareness (TIA) project mentioned above, Ryan McKinley, while a student at the MIT Media Lab, developed Open Government Information Awareness (http://opengov.media.mit.edu, now defunct), a website that collated information about government actors (http://web.mit.edu/newsoffice/2003/gia.html) , such as members of Congress, via public sources (Congressional records, C-SPAN video archives and transcripts, and so on). The idea was to allow members of the public to watch members of Congress (http://www.sourcewatch.org/index.php?title=Government_Information_Awareness) , just as the government was threatening to do with TIA. The project unfortunately did not last long due to similar media outcry and publicity.
Other, less confrontational data mining works have additionally re-purposed these techniques to produce radical cartographies of governments and multinational corporations. They Rule (http://www.theyrule.net/) is one such project that enables people to make links between the Boards of Directors the largest multinationals, showing how numerous directors are present on multiple boards. Josh On, one of the main people behind the project, wrote about how he hoped that projects like his would help shift the configuration of power from ?They? to ?We?: ?But as artists, as people, we should not shy away from the most important task that confronts us, organizing collectively to achieve a world in which we can honestly say: We rule. (http://220.127.116.11
/en/archiv_files/20021/E2002_367.pdf) ?. Additionally, the French group bureau d?études (http://bureaudetudes.free.fr/) developed a number of maps that laid out the relationships (http://bureaudetudes.free.fr/act.html) between agencies of the US Government as well as non-governmental organizations such as OPEC.
Recently the group RYBN (http://rybn.org/) has been creating a series of projects under a rubric of what they call antidatamining
(http://antidatamining.net/) . This work aims to use the techniques of data mining to present information about capital and financial institutions in alternative ways to those traditionally seen in business contexts. Their most recent work, Stock Overflow (http://www.imal.org/StockOverflow/) extends this into a series of installations and conversations with media theorists about how to ?recontextualize? the present financial ?crisis?.
Avoiding the void of despair
As we begin to make-the-links-that-we-were-not-supposed-to-make, it becomes easy to see the mass as a void that would engulf us in a clingy, clammy, sticky despair. Diving further in we become swallowed in the web of edges that connect the nodes of the powerful actors together, unable to move, still, in a paralyzed stasis?
Yet what we need instead is movement, the proverbial (by now) ?lines of Right? that, yes, make these links visible, but additionally break them open through acts of resistance, of the unexpected link to an alternative network. We need the playful-serious, the ?-? ever more important as the link we make ourselves between that-which-we-must-find-out and that-which-we-want-instead. This is our own version of Guattari?s ethico-aesthetic paradigm, the development and growing of our own individual and collective subjectivities that do not deny the gravity of the situation but do deny the ability of the situation to have complete hold over us. Libre software, radical street bands, TAZs, immediatism, alternative currencies, pirate radio, hacktivism, guerilla gardening, circuit bending, freeskools, street dances; what we want is not new (http://www.radicalsoftware.org/e/index.html) , yet we need to continually reactivate it in the face of sustained counter resistance. ?Poetic austerity? is our term du jour?and it will probably change tomorrow. While we fight for the state to provide more than the basic conditions of existence for everyone, we simultaneously (re-)construct alternative and parallel forms of pedagogy, exchange, and communion. We consider poetic responses based on our present abilities, absent the support structures we hope one day to have. Our practices recognize our materials and support around us: from the cast-offs at the thrift-store, to the colleagues around the world who share their programs. Call us naive, if you will; we don?t care. The alternative power of conjoining the poetic and serious enables us to not only respond to this so-called ?crisis? (a regular event within the history of capitalism) but also to ferment our own links and develop our own tactics and (gasp!) strategies that recognize our present material position while not being limited by it.
Appendix: Technical Details
The technical details regarding the sources that MAICgregator searches can be found in the frequently asked questions (/FAQ) and examples of the output can be seen in the screenshots (/docs/screenshots) . More specifically, the extension talks to a MAICgregator server that runs the code that performs the searches that aggregates the data. This code will shortly be made available so that anyone can run a local installation of the server-side application. We store the results in a variety of database types, including SQL, XML, and RDF, depending on the type of data retrieved. For a sneak-peak of what this all involves, please see the basic server installation instructions (http://dev.maicgregator.org/wiki/Server%20Installation) . Aspects of this database will be made public shortly, especially the growing network of trustee relationships that we are building. We are considering how to release the rest of the database; given our own esoteric means of saving and retrieving the data we get, it will be difficult for others to use the data exactly as we have it. But perhaps that is okay; perhaps all is necessary is for this aggregate to be available, no matter the difficulty of looking inside.