[-empyre-] Welcome to Week 3 on -empyre-: Internet/Online Representation
everett at filmandmedia.ucsb.edu
Tue Apr 21 05:32:47 AEST 2015
Aphid, et al,
I am enjoying our panel's exchanges between ourselves, and hope others
will join in at some point before Wednesday. Anyway, following up on
your interest to hear more about what it is in recommender algorithms
that I am contemplating, it has to do with how I feel profiled by
these big data and data mining processes. So, what got me thinking,
and I may need you and the other co-panelists to talk me down here,
why are nearly all e-commerce ads being pushed to me featuring black
women models? I began noticing that occurring in print catalogues from
department stores. It is as though there is the black and white
consumer versions of ads. Are there Asian and Latina/o ones as well? I
do not recall being issued a census or identity block to check when I
made my purchases online.
So, in anthropological analyses, I know there is a caution about how
survey respondents often deliberately skew results by misreporting,
etc. In that vein, I am often tempted to email Macy's or other online
retailers claiming that I am NOT black and am offended by this
marketing in an effort to see if there is a way to get at what you ask
about the proprietary nature of these application programming
interfaces (API) or identity seeking bots as I prefer to call them in
the case of my retail shopping profiling. In the context of Amazon's
recommender algorithms for book buying (they were the first to really
massify this technology, I think), I really appreciate that data
mining and recommending. It is useful when I am looking for new
scholarly books in a certain area or discipline. But, in the case on
online shopping, on social media ads targeting, this seems a bit
spooky to me. Of course, this means that I will try to do some type of
analysis of this process after I complete the other overdue projects
with which I am currently engaged. Am I being unduly concerned?
In following your embedded links, I was struck by the photo
accompanying the now-defunct Face++ (bought out by Facebook, as you
point out). On that site's "about us" page, there is an image of an
all-white group of folks seemingly jumping for joy about something
(maybe the absence of people of color (POC); just sayin,' here is the
link to that photo <http://www.faceplusplus.com/about/>.
I am concerned because for me there are issues of identity mapping and
profiling that obtain in these backbone software processes and
protocols. And they matter because, even though it may be true that
identity politics matters and contemporary communications studies are
perceived as incompatible, per Soraya's point earlier, I think they
must not be. I mean, what about the fact that Ben Affleck tries to
scrub the historical fact that his ancestors were slave-holders? And
what about this story--if true--"Missouri town elects first black
mayor; 80% of police, city attorney, water supervisor resign." Come
on, if scholars are interested in or do not find such identity matters
of import/consequence, then I thank god for those of us humanities
scholars who do.
On Sat, Apr 18, 2015 at 7:58 PM, abram stern (aphid) <aphid at ucsc.edu> wrote:
> ----------empyre- soft-skinned space----------------------
> Hi fellow panelists and empyrians,
> Trying to weave together (as I think we all are) some of the strands of research and conversation with some stuff I've had stewing here.
> I love this question: "What are the ideological underpinnings and agendas at work?" -- to turn this into a methodological question, how can we expose the ideological underpinnings and agendas at work when the technologies and the institutions that deploy them are often so opaque? For instance, you mentioned research on recommender algorithms (eager to hear more on this!). Is there any transparency as to how they operate? One of the difficulties in analyzing much of this technology is that it is proprietary; much of the apparatus and its algorithms are obfuscated.
> One of the artworks I found when doing research on the face detection Sanctum is an art project produced by Juan Pampin and James Coupe currently being exhibited at the University of Washington's Henry Art Gallery; as you approach it it detects your gender and approximate and age using a commercial API* and then shows you images and social network updates of people 'like you' on screens; in speaking with the artists, I learned that it records gender on a 30 point scale, -15 being the most masculine and +15 being the most feminine. Is this non-binary gender coding (by the API designers [or perhaps by whatever researchers they based the API on], not the artists) an assertion of progressivity or addressing a technical limitation of a system that can only guess its subject based on similarities between two defined poles produced by compositing lots of 'male' and 'female' faces? Both answers suggest underpinnings but they go off in different directions.
> One last thing I'll leave here since it seems on topic:
> http://aphid.org/ucsc_ethnicities.jpg I took this photograph outside of an administration office here at UCSC, it's part of a larger poster that shows how students are stored within the University's IT system. Yes, that's a very 90's matrix-y face in the background!
> More soon,
> *I don't know which API - http://www.faceplusplus.com/detection_detect/ is an example, check out the JSON about halfway down and imagine it being run across social networks, even if only on 'public' photos. Yikes. For what it's worth, Face.com was very popular API for face matching (a colleague used it to identify members of the VA senate in an archive of their legislature) until a couple years ago when facebook bought them up and closed down the service.
> On Sat, Apr 18, 2015 at 8:25 AM, Anna Everett <everett at filmandmedia.ucsb.edu> wrote:
>> ----------empyre- soft-skinned space----------------------
>> Aphid, your research into the politics of the archive is quite
>> exciting. And in particular I find your ideas about how archival
>> "records provide opportunities to re-present, reconfigure and detourn
>> the performance of politics into more critical engagements with issues
>> these institutions seem unable to address on their own." In many ways,
>> your work transects some of my own, some of Kishonna's, and some of
>> Andre's, especially your current project on the "Unreliable
>> Interrogator." I am pleased that all of our works are attuned to the
>> ideological basis of who may read, write and archive content, in your
>> analysis, who controls diversity narratives in Kishonna's, and what's
>> next in technocultural politics of respectability in Andre's?
>> Underlying all these is the question of what is at stake here? What
>> are the ideological underpinnings and agendas at work.
>> Aphid, when you said "I am deploying visual, auditory and textual
>> analysis on records of the SSCI. One mode utilizes facial feature
>> detection, locating a speaker's face in video through which various
>> demographic information can be derived (race, gender, age -- very
>> unreliably) and compared . . ." it reminded me of a current line of
>> inquiry that I am working up. That is interrogating the
>> intersectionality of big data's recommender algorithms and race. I
>> will discuss that a bit more more in my next post. Thanks for
>> reminding me of that.
>> Fellow panelists, I am so pleased to be in this discussion topic with
>> you alll. I am wondering, though, where are the rest of the
>> empyrians??? I know you guys have something great to say. I know
>> because lurked last week. I could not participate because I was
>> traveling and wrapping up an overdue project.
>> Anna .
>> On Thu, Apr 16, 2015 at 2:55 AM, abram stern (aphid) <aphid at ucsc.edu> wrote:
>> > ----------empyre- soft-skinned space----------------------
>> > Hi all, excited to be part of the discussion here. Thanks, Soraya, for the invitation!
>> > My work primarily addresses the politics of the archive, particularly looking at how the material of public media (in particular, government-produced, public domain documents) are produced/reproduced and distributed. I'm interested in the way in which democratic institutions present themselves, through online archives and collections, to their publics. Most of my projects have been implemented as prototypes, trying to imagine how open and participatory archives might operate in the same institutional context as what I critique. I'm increasingly interested in how these records provide opportunities to re-present, reconfigure and detourn the performance of politics into more critical engagements with issues these institutions seem unable to address on their own.
>> > I'm about a year into my current project, The Unreliable Interrogator, which is still very much a work in progress. It's a return to internet art for me after a decade of more traditional research projects and prototypes. The Interrogator's focus is the online Hearing Archive of US Senate Select Committee on Intelligence (SSCI). The SSCI is the main oversight body of the US intelligence apparatus (oversight denoting both the act of overseeing and a failure to notice). The Committee's website, which still bears a Copyright notice from 2006, only lists hearings from 2009 forward despite containing hearing data back to 2002 (if you know where to look!). Hearing videos exist, but hide behind broken and absent links.
>> > The first part of this project is a process of media archaeology and interrogation: locating and downloading videos, testimony PDFs and associated metadata (witness names, titles and so on) and converting, when necessary, to more legible or open formats before publishing to the Internet Archive. This excavation serves as a platform -- for media reuse but also for critique: an interrogation of mediality, the formats deployed by the committees are also products of ideology and deserve scrutiny for the way in which they delimit who may read, write and archive content through the use of Digital Rights Management, network throttling and other methods.
>> > The second part of the project uses the reconstituted archive discussed above as a platform for a different mode of media interrogation. Using widely available tools and public APIs, I am deploying visual, auditory and textual analysis on records of the SSCI. One mode utilizes facial feature detection, locating a speaker's face in video through which various demographic information can be derived (race, gender, age -- very unreliably) and compared against witness and committee composition data (all white, until this year when Hirono joined the Committee). Another visual analytic mode uses cut detection to determine when camera changes occur, which also signal a change in speaker, allowing for a rough mapping of the discursive shape of a hearing. An audio analytic searches for content that won't occur in a transcript, off-mic chuckles, breathing, and silence. Testimony PDF text is compared against a list of terms the Department of Homeland Security apparently uses to mon
>> itor social networks for threats.
>> > This "work" is distributed, and takes place in the web browser of each visitor for as long as they let it run, in a way like SETI @ HOME (I suppose both seek intelligence in unlikely locations /rimshot); data gathered is broadcast to a server which will collect and publish reports nightly. I'm reluctant to link to things that barely or don't work, but here's a prototype of the audio analysis module that should run in an up to date chrome or firefox: http://unreliable.interrogator.us/sound/ht.html
>> > In terms of key challenges, I've been stuck on something Trevor Paglen said at UC Berkeley last year at an event on Pan-Optics: (paraphrasing; the talk wasn't recorded) "Representational media is slowly being replaced by operational media, which is made for and by machines with us as its subjects and targets."
>> > There's more, but this is already running longer (and later!) than I intended. Looking forward to the discussion!
>> > Best regards,
>> > Aphid
>> > On Wed, Apr 15, 2015 at 9:41 PM, Andre Brock <brocka at umich.edu> wrote:
>> >> ----------empyre- soft-skinned space----------------------
>> >> Hello, all!
>> >> Thank you for inviting me to this discussion, and "hey!" to my co-panelists this week!
>> >> My intellectual investment in this area has changed over the last decade. When I first started researching online Black identities, my intention was to follow the path blazed by folk like Dr. Everett, Alondra Nelson, and Lisa Nakamura - to highlight that minorities (in my case, African Americans) were always already part of the burgeoning online spaces hailed as Web 2.0. I felt then, as i do now, that it's important to interrogate digital interfaces for the meanings their designers impose, while also mining discourses created by users of those interfaces to see how they adapt themselves to the new medium. Over the last few years, i've examined browsers, video games, and Twitter to see how digital representations of Black identity have been composed, contested, and destroyed.
>> >> Given the proliferation of attention towards Black digital activity over the last few years, from the Commander-in-Chief on down to Black Twitter and even back to HotGhettoMess.com, I've become interested in a different take on Black online representation.
>> >> As more and more Blacks access online spaces, segments of Black online users have begun to articulate a growing technocultural politics of respectability. The "New" New Blacks (h/t Ms. Anna) are much more interested in political and polite representations of Blackness online, and in the process deprecate the gossip blogs, ugly BlackPlanet pages, and ratchet-ass tweets and images that signaled the spread of Black culture to online venues. My current work has me investigating the way that perceptions and arguments about Black Twitter use have shifted over the last few years - from celebrations of folk culture and 'individual' blackness to an insistence on 'proper' political uses of the service.
>> >> I'm especially pleased to be part of this discussion, as I have of late been pondering my role/stake as an internet researcher pursuing questions of African American identity in various digital spaces. I have three questions about identity, representation, and the digital/Internet that are pulling in me in multiple directions, and it is my hope that my articulation of them can push this conversation in interesting directions.
>> >> 1) Whither identity in post-PC, post-racial (heh) cyberculture and the continual fragmentation of online representation? Are we our devices? Our wearables? Our profiles?
>> >> 2) Isn't it past time that research into performances of identity and online representation start addressing Whiteness? Particularly in the light of GamerGate, MRAs, and #alllivesmatter, where is the corresponding research into how racial ideologies shape White online identity, especially since the aforementioned online movements draw heavily upon digitally-mediated beliefs about race and gender?
>> >> Brenda Laurel's excellent commentary last week referenced race and GamerGate, but I am always troubled when deviant activities are assumed to be perpetrated by deviant society members. Brenda's (if i may be so bold) argument that GamerGaters are men with "poor educations and a degree of poverty" seems to let 'brogrammers', Kleiner Perkins VCs, and other highly educated wealthy white (and non-whites afflicted by false consciousness (yeah, I said it) men and women off the hook. Just like Klansmen were usually highly respected businessmen in their communities, much of the racist and sexist online activities we see daily are done by elites, not just by the dispossessed.
>> >> *Kishonna, i know that your work addresses race and gender specifically WRT gaming, but (correct me if i'm wrong) not many folk are making connections between XBL gamer behavior and GamerGater behavior.*
>> >> 3) With the maturation of minority political activism in social networks (#blacklivesmatter) and near parity in material access/broadband access through mobile devices, is it time for new media/internet research to move past online identity politics and online representation? /sarcasm
>> >> If so, where do we go from here?
>> >> That's all i have for now...i look forward to seeing what's on the minds of my co-panelists and the empyre audience. Thanks for having me!
>> >> PS - Ms. Anna, congrats on the new publication!
>> >> _______________________________________________
>> >> empyre forum
>> >> empyre at lists.artdesign.unsw.edu.au
>> >> http://empyre.library.cornell.edu
>> > _______________________________________________
>> > empyre forum
>> > empyre at lists.artdesign.unsw.edu.au
>> > http://empyre.library.cornell.edu
>> Anna Everett, Ph.D.
>> Department of Film and Media Studies
>> 2322 Social Sciences and Media Studies (SSMS) Bldg.
>> University of California
>> Santa Barbara, CA 93106-4010
>> Email: everett at filmandmedia.ucsb.edu
>> Phone: 805.893.8706
>> Fax: 805 893.8630
>> empyre forum
>> empyre at lists.artdesign.unsw.edu.au
> empyre forum
> empyre at lists.artdesign.unsw.edu.au
Anna Everett, Ph.D.
Department of Film and Media Studies
2322 Social Sciences and Media Studies (SSMS) Bldg.
University of California
Santa Barbara, CA 93106-4010
Email: everett at filmandmedia.ucsb.edu
Fax: 805 893.8630
More information about the empyre