back to

e-flux journal: Hito Steyerl on "proxy politics"


Hacer24580318′s Twitter feed advertises Flappy Tayyip, a game starring Turkish president Tayyip Recep Erdoğan.

In the latest issue of e-flux journal, Hito Steyerl plumbs the depths of cell-phone photography, dirty pictures, and what she calls “proxy politics”:

To better understand proxy politics, we could start by drawing up a checklist:

Does your camera decide what appears in your photographs?
Does it go off when you smile?
And will it fire in a next step if you don’t?
Do underpaid outsourced IT workers in BRIC countries manage your pictures of breastfeeds and decapitations on your social media feeds? bottle
Is Elizabeth Taylor tweeting your work?
Are some of your other fans bots who decided to classify your work as urinary mature porn?
Are some of these bots busily enumerating the names of nation states alongside bodily orifices?

Is your total result something like this?

(’I*) (*’σ з`) ~♪ (*’台`*) (*≧∀≦*) (*゚ェ゚*) (*ノ∀)
) (/ε\) (/ε\)

Congratulations! Welcome to the age of proxy politics!

Read the rest of this astute and offbeat essay here.


hey did anyone else notice that the entire premise of this article (how cell phone cameras work) is like not rooted in any verifiable reality?

good read though, thanks!


Here are some example of papers describing research similar to the one mentioned in my text, ie linking face recognition algorithms on mobile device cameras with social media data bases. This was widely discussed also in the context of googles acquisition of viewdle and fb´s acquisition of, apps like recongizr, name tag etc, and of course facebooks deep face technology Literature on different methods of denoising cell phone pictures abounds. On averaged probabilistic faces see discussions around eigenfaces vs fisherfaces.

4.3. Face Recognition using Social Context Recent research has shown that social context information can improve face recognition accuracy. The datacrawled from Facebook provides us with such context in-formation, and in order to include it, we follow an approachproposed by Stoneet al. [21]. For each query, a pairwisey1y2y3p(y3|x)p(y1|x)p(y2|x)p(y1,y3|x)p(y1,y3|x)p(y1,y3|x)Figure 5. Example of a Conditional Random Field. The facesare represented through nodes and the edges reflect the relation-ships [22].conditional random fieldG= (V;E)is defined, where eachnodev2Vrepresents a face and the edgese2Erepresentthe relationship between each pair of faces. An illustrationof such a graphical model is shown in Figure 5. Also, thepossible discrete label spaceL=fl0;::;lNgis defined. Inour caseLcontains all friends of the user. Given an inputimage accompanied by meta dataxwe try to map face iden-titiesy=fyi2Lgto all nodesiin the graph. Stoneet althen represents the model asf(y;x) =Xii(yijx) +X(i;j);i6=jij(yi;yjjx)(5)withi(yijx) =Xmm(x)gm(yi;x)(6)i(yi;yjjx) =Xnn(x)g0n(yi;yj;x)(7)The mapping functionfis a linear combination of differentunivariate and bivariate feature functionsgmandg0n. Thelearned weightsmandnare used to combine the featurefunctions.In our implementation we use two different univariatefeature functions similar to [15, 26, 21]:appeareance(yijx) =1faceScore(i;x)(8)popularity(yijx) =2Pp2P(yi;p)jPj(9)(yi;p) =(1personyiappeare in imagep0else(10)where1and2represent two optional normalization con-stants andPis the total set of images available throw thenetwork. The functionfaceScorerepresents the confi-dence of the k-NN classifier described in the previous sec-tion. Naamanet al. [15] observed that some people tendto appear more often in private photo collections then otherpeople. They also observed that people that appear togetherin the same image, have a higher probability of appearingtogether in other images. The first observation we describewith the feature functionpopularityand the second onewith the bivariate feature functioncooccurrence.cooccurrence(yi;yjjx) =w3Pp2P(yi;yj;p)jPj(11)(yi;yj;p) =(1yiandyjappear together in imagep0


Computational Photography doesn’t depend on my network. It can " scan all other pictures stored on the phone or on your social media networks and sifts through your contacts", but that’s a very narrow definition of what it actually is and how it works. That’s just one way to do it; a very messy way to do so, if I may say so. What if I don’t have any social media presence? Since when did I gave permission to the camera do so? This is would be a total nightmare for a company like Apple. The association between computational photography and a previous database of images gathered by the user and his friends is terribly misleading. This just in: “Apple scans your network photos when you take a picure.” Nope, sorry, not happening, not now, not EVER. There’s not a single manufacturer that does this out of the box, by the way. Apple does use algorithms that MAY fall within the definition of computational photography but they just don’t work like that. Like, at all.

What the article describes seems to be a very specific app, from a very specific developer, that for some strange reason never goes mentioned.

It’s one thing to say that a specific app can do that, but it’s just wrong saying that’s how modern cell phone cameras work.


@hito thank you for posting that article! it was very interesting, and i’m sure many phone cameras will operate that way in the future. however, in the section directly after the one you quoted, the paper explicitly states the drawbacks of that functionality and how it’s as of yet impractical to implement broadly in cell phones. the way you present face recognition and computational photography in your article is misleading, if not untrue (at least for now). while there certainly are variants of this implementation on the app-level as @elfmachine stated (i say this after noticing skew lines becoming straight once posted to instagram) the actual photos taken pre-upload are largely free from meddling algorithm-fingers that access networked information and are, in fact, just shitty pictures taken on a crappy camera.

it’s also worth noting that the program in the article you linked to actually does nothing to intervene with the content of the photo itself. while such technology certainly will and may already exist, the complete system you describe in your article is not one that many (or any) users have access to. all this is to say that these things are possible, and when they become pervasive we will certainly have an entirely new set of worrying circumstances to navigate, but to assert that they are a part of the reality we are living in is a bit prematurely apocalyptic.

i only point this out because i found your article genuinely interesting, and i felt it contained some very prescient points that are worth delving into. a lot of the conclusions you make still stand even when altering the computational photography premise to similarly obfuscating and potentially oppressive ways photographic technology is more commonly used right now (i’m thinking more along the lines of how photos exist in databases, and the narrativizing power of metadata).


simle :slight_smile: it’s autoawesome :stuck_out_tongue: