Image Recognition

Katrina Sluis

How might we consider exhibitionism in a media landscape characterised by pervasive visibility and enduring exposure? From drones to cellphones, photography is now ubiquitous and shared with a casual click; our movements online are tracked, analysed and fed into algorithms that in turn filter our movements through the network. Where once we mastered HTML to get online, we now worry that online platforms now master us. Today, our Linkedin profiles harass us until they are 100% complete, Facebook wants us to identify which friends are actually siblings, Twitter wants to know our thoughts and YouTube asks us to Broadcast ourselves. The data that forms as a by-product of our sharing and liking has limited value unless its provenance is assured, and has been linked to a verified human (whether attribution is coveted or not).

Geert Lovink has argued that whilst “[e]arly cyberculture was driven by a shared desire to be someone else” the second Internet bubble relies on the loss of anonymity: the desire to be seen, to disclose, to like and be liked [1]. But as our newsfeeds warn us repeatedly, cyberspace is full of danger and populated by spammers, child molesters, hackers and trolls, ready to steal your password or impersonate you. Automated bots are some of the most prolific users of Twitter, and it is estimated that 51% of Internet traffic is non-human and  ‘mostly malicious’ [2]. Lovink notes that in a post 9/11 world in which hysteria about security has intensified, web users have taken sanctuary in the safe filtered citadels of Web 2.0 platforms [3]. Facebook creates a cocooned visual world, protecting its users from hackers who spread pornography, shielding the community from images of breastfeeding women and amateur drawings of nudes [4]. In order to “help fight spam and prevent fake profiles” Google has established “community standards” allowing it to repeatedly block the accounts of users who persist in using a pseudonym [5]. Sheltered in a proprietary social network, total exposure is not such a threat: we are among friends, our audience is known, and we can choreograph our performance accordingly.

Faced with the labour of maintaining multiple online profiles, it becomes only logical to connect one’s Twitter account to one’s Storify account to one’s Facebook account to one’s Instagram account to one’s Yahoo account – so convenient and neat! In the post-Facebook world, Mark Zuckerberg has famously proclaimed: "You have one identity… The days of you having a different image for your work friends or co-workers and for the other people you know are probably coming to an end pretty quickly… Having two identities for yourself is an example of a lack of integrity." [6]

What Zuckerberg here fails to disclose is that the consolidation of fragmented and multiple online identities into a single verifiable person have become central to an Internet economy that extracts value from every mouse click. Or, as Schlapa from Pussy Riot suggests: "Capitalism is based on the principle ‘buy and sell’. This only works when you have a face and show it. A person without a face does not appear as a market participant. Capitalism does not tolerate anonymity." [7]

Giving up one’s anonymity is increasingly a necessary precondition for social participation, under technological conditions in which security and marketing progressively go hand in hand. This has become possible because, as Zygmunt Bauman suggests, ‘[t]he condition of being watched and seen has … been reclassified from a menace into a temptation’ [8]. Bauman argues that making one’s life a matter of public record, and exposing oneself online has become an irresistible temptation for a culture where the threat of exclusion has largely replaced the threat of confinement: "On the one hand, the old panoptical stratagem (‘you should never know when you’re being watched in the flesh and so never be unwatched in your mind’) is being … brought to well-nigh universal implementation. On the other, with the old panoptical nightmare (‘I am never on my own’) now recast into the hope of ‘never again being alone’ … the fear of disclosure has been stifled by the joy of being noticed." [9]

Social media platforms manifest this paradox by combining visibility and permanent exposure with the continuous threat of invisibility [10]. The relentless realtime scroll of contemporary communications demands continuous participation to ensure one remains visible ‘above the fold’ and to guard against one’s own obsolescence. Providing the tools of self-fabrication, social platforms offer the opportunity to broadcast who one is connected to, exhibit who one is loved by and demonstrate proof of social recognition [11]. The contemporary desire for hyper-recognition is, for Gilles Lipovetsky, closely linked to a culture of ‘hyperindividualism’ in which “one’s sense of identity and belonging is anything but instantaneous, given once and for all.” Instead, identity has become “open and reflexive, an individual gamble in which the dice can be thrown again and again.”[12]

The Quest to be Recognised

As a technology specifically of visuality, photography operates both in the register of exhibitionism and spectacle, by having both a one-to-one relationship with the body on the one hand and the commodity image on the other. It holds a privileged place within the Internet’s economy of attention, by channelling interest toward specific narratives, generating fame, money or notoriety for their distributor-curator-authors. Images wind their way past spam filters and into mailboxes, containing seductive promises of riches, miracle pills and other cosmetic delights. It is alleged that stories posted to Facebook with an image generate 54% more attention [13]; with the right headshot and Google+ profile the content you create can attract anywhere from 40% to 150% more clickthroughs [14]. Where once malignant users might have ‘phished’ for bank account details through bogus email requests, today Facebook’s ‘like’ button is the preferred tool for accumulating the riches of the network. ‘Phishing’ now exploits the affective power of the networked snapshot, and takes the form of an image of a small child on Facebook, holding up a sign, asking for the ‘thumbs up’ if you agree they are beautiful or deserve medical treatment. Having accumulated thousands of ‘likes’ the page can then be auctioned off to marketers seeking to profit from the Internet’s attention economy.

Exclusion and visibility on social platforms is intimately linked to the arrangement and design of software itself, by modulating attention and the ‘performative infrastructures’ in which participation takes place [15]. For the creative consumer, the key to standing out involves gaming the appropriate algorithm and understanding the particular affordances of traffic. The blogosphere is full of stories of lucky individuals in the right place at the right time with the right cameraphone who sold their images via Twitter. With the right eccentric-looking-yet-photogenic cat, we are told, we can join other celebrity cat owners in joining YouTube’s affiliate programme and never have to work again. On Flickr, amateur photographers swap tips on how to game the ‘Interestingness’ algorithm; coalitions of users form to generate tags, distribute awards and probe the interface in order to find ways to enhance their social currency [16]. Professional photographers now learn the nuances of Search Engine Optimisation in order to monetise the Image Search traffic driving visitors to their site. Artists and job seekers alike are encouraged to develop a tweet strategy to create a personal brand, with each 160 character missive composed and synchronised to the global time zones of their distributed target audience.

 Bounce Rate

But as recent reports suggest, it is only a minority users who participate in ‘upload fever’ by authoring and sharing their own original content [17]. Rather than a creative utopia, we have Tumblr. Just as photoblogging offered a casual and less time consuming alternative to blogging; Instagram’s filters have liberated photographers from the labour of conceptualising a unique aesthetic. Platforms such as Pinterest and Tumblr provide the ideal tools with which to craft an online persona and validate one’s Internet connoisseurship through image consumption. However, because these platforms are characterised by the continuous re-circulation of images, there is a premium on original content. Brad Troemel observes that in the relentless game of attention on Tumblr, the key is to become the cited origin of an image which has been repeatedly shared, because, “[i]t isn’t as impressive if a post has 10,000 notes and you’re the 10,001st person to reblog it.” [18] Here, Troemel notes, it is not the glory of being seen as the creator of the image, but curator – a network user of impeccable taste with a talent for drawing attention to the most abnormal, interesting or obscure content.

Lost in the sensuous plenitude of images which one can link to, like, reblog and retweet, a range of tools have emerged to give us the ability to assess the value of our ‘brand’ and the images we share. What tag to optimise for? What images re-posted attracted the most views or notes? How long did a visitor dwell on a certain post? What content sent them elsewhere and from which site did they travel? The disappearance of ‘photographs’ into ‘content’ here is paralleled by the transformation of the ‘audience’ into ‘traffic’. Jodi Dean suggests that the provision of tools for counting, measuring and imagining our audience is indicative of the fact that our spectators are, in the end, ultimately unknown to us: we may imagine we are amongst friends, but ultimately we perform to an obscure or unknown point in the network [19]. Or, perhaps, to ourselves. As contemporary social life becomes characterised by hyperindividualism and hypersurveillance [20], Dean argues the seduction of social network sites is clear: “[they] let us see ourselves being seen” [21]. In a society of hyperconsumption, Bauman argues, we become consumers of our own image and marketers of ourselves as commodities: “that is, as products capable of drawing attention, and attracting demand and customers.” [22]


Retreat from Representation?

But in an age of pervasive media, Hito Steyrl notes that one consequence of the ubiquity of images, is that “pictorial representation – which was seen as a prerogative and a political privilege for a long time – feels more like a threat”. In an age of endless representations, endless self-fashioning, there emerges another, contrary desire: “to be invisible, if only for 15 minutes” [23]. According to the New York Times, it took the photographer Rich Lam a single day to discover the identities of a mystery couple caught kissing in one of his photos taken during street riots in Vancouver [24]. New Scientist reports that it is now possible for facial recognition systems to identify a single face from a database of 1.6 million photos within 1.2 seconds with 92% accuracy [25]. And if you are lucky enough to survive your teenage years without having a traumatic experience photographed and shared online, there is still the risk you could be “Facejacked” and reanimated as Steven, 44 from Melbourne, appearing in Facebook ads to tell your friends how they, too can work from home and earn that dream salary they deserve [26].

In their paper, Achieving Anonymity Against Major Face Recognition Algorithms, Benedict Driessen and Markus Dürmuth describe an algorithm which supports ‘face de-identification’ to prevent automatic recognition by machines, whilst ensuring the photographic portrait remains recognisable to a human audience. This level of anonymity, they noted, was important to protect users of social networks against two pressing issues: “First, automated extraction of the social graph, i.e, friendship relations; second, automated tagging of people in images” [27]. This becomes a pressing cultural issue, because whilst early forms of communities appeared transient, the centralisation and enclosure of the Internet makes it clear “the Internet never forgets” [28]. In a 2010 interview with The Wall Street Journal, Google’s Eric Schmidt predicted that one day there will arise the need to give young people the automatic ability to change their name by deed poll at adulthood in order to escape their adolescent selves archived and available on their friends’ social media sites [29]. But the question remains what will happen when facial recognition technology advances to the point where any orphaned ‘selfie’ can be discovered and linked back to a concrete identity and Facebook profile? Will a name change be enough to escape one’s online history once facial recognition technologies become ubiquitous?

Noting the way in which the networked cameras of today have the potential to ‘drain away your life’, Steryl describes a pattern in which a number of groups now actively resist being photographed, interviewed, monitored, tracked or represented [30]. Whilst Steryl focuses on examples where protesters have smashed cameras or deliberately concealed their bodies, there is a parallel trend online where users are adopting tools to camouflage, obfuscate or anonymise their presence online. Finn Brunton and Helen Nissenbaum provide examples of Facecloak, TrackMeNot, BitTorrent Hydra and CacheCloak as examples of ‘vernacular resistance’ to forms of tracking and monitoring [31]. As tools of misrepresentation, they work to mutate the shape of one’s data shadow through the generation and submission of fake information on one’s behalf. For the cypherpunk community, independence is to be sought ‘behind a cryptographic veil’, and the development and dissemination of software to facilitate anonymity online through encrypted communications. [32] And as Facebook fatigue begins to take hold, a range of libre and peer to peer platforms are emerging which attempt to imagine alternative models of sociality and interfaces to support this.

There is of course an argument in which total transparency can become a form of ‘empowering exhibitionism’ in which ‘people seek to play an active role in the production of images, thus, reclaiming the copyright of their own lives’ [33].

However, contemporary modalities of visibility and exposure cannot be considered outside of the technosocial platforms through which our digital doppelgängers are presented, arranged and valorised. The pressing question is rather, what kinds of subversion becomes possible under the framework of communicative capitalism through which, Dean asserts, “contemporary communications media capture their users in intensive and extensive networks of enjoyment, production and surveillance.” [34] They key problem here lies in the fact that contemporary subjectivity is increasingly bound to processes of self-exposure and reflexivity which today “captures creativity and resistance so as to enrich the few as it placates and diverts the many.” [35]

This is not to suggest that social softwarebecomes all encompassing and that network participants are unable to play with or subvert the interfaces they perform with. However, such platforms reveal the need for a more nuanced consideration of the meaning of participation, alongside other traditional dichotomies such as inside/outside, public/private and commercial/political, democratising/disempowering [36]. Furthermore, the intensification of affect as a byproduct of algorithmic recirculation and repetition online calls for a shift in understanding of the photographic image, as Carlos Barreneche suggests, in which images are considered   “not as carriers of meaning but rather as informational objects that in their dissemination (flow) acquire qualities that are susceptible of calculation.”[37] Rather, it requires close attention to the image’s doubling as representation and information and “the operational logics governing the access to whatever media object and ultimately frame the production of meaning as such”. [38]


  1. Geert Lovink. Networks Without a Cause: A Critique of Social Media. Cambridge 2011, 39. back
  2. Tom Forenski ‘Report: 51% of Web Site Traffic Is “Non-human” and Mostly Malicious’  ZDNet. March 13, 2012 Accessed June 2, 2013. http://www.zdnet.com/blog/foremski/report-51-of-web-site-traffic-is-non-human-and-mostly-malicious/2201. back
  3. Lovink, Networks without a Cause, 40. back
  4. See: ‘Facebook Hit with Unsolicited Porn, Violent Videos,’ Firstpost, 16 Nov 2011, Accessed June 3, 2013. http://www.firstpost.com/tech/facebook-hit-with-unsolicited-porn-violent-videos-131970.html. And also: Charles Arthur, ‘Facebook’s Nudity and Violence Guidelines Are Laid Bare,’ The Guardian, February 22, 2012. Accessed June 2, 2013. http://www.guardian.co.uk/technology/2012/feb/21/facebook-nudity-violence-censorship-guidelines back
  5. Marcel Rosenbach and Hilmar Schmundt, ‘Internet Evolution: The War on Web Anonymit,’ Spiegel Online, 8 May 2011. Accessed June 2, 2013.  http://www.spiegel.de/international/spiegel/internet-evolution-the-war-on-web-anonymity-a-778138.html back
  6. Mark Zuckerberg in David Kirkpatrick, The Facebook Effect: The Real Inside Story of Mark Zuckerberg and the World’s Fastest Growing Company. New York, 2010, 199. back
  7. cited in Oliver Leistert ‘Smell the fish: Digital Disneyland and the right to oblivion’, 
First Monday, 18: 3, March 2013
  8. Zygmunt Bauman and David Lyon, Liquid Surveillance: A conversation, Cambridge, 2013, 23 back
  9. Bauman and Lyon, Liquid Surveillance, 23 back
  10. This idea is developed in detail in Taina Bucher, ‘Want to be on the top? Algorithmic power and the threat of invisibility on Facebook’, New Media & Society 14: 7, 1164-1180 back
  11. Bauman and Lyon, Liquid Surveillance, 34 back
  12. Gilles Lipovetsky, Hypermodern Times, Cambridge, 2005, 65. back
  13. Peter Kafka ‘Want My Attention on Facebook? Post a Picture on Friday Morning.’ AllThingsD. Accessed June 3, 2013. http://allthingsd.com/20100921/want-my-attention-on-facebook-post-a-picture-on-friday-morning/. back
  14. The blogs of marketing and SEO specialists abounds with various claims concerning clickthroughs, such as “How Optimizing My Ugly Google+ Pic Increased Free Traffic 35%.” SEOmoz. Accessed June 4, 2013. http://moz.com/blog/google-author-photos and “How Rich Snippets Can Improve Your CTR | Catalyst.” Catalyst. Accessed June 4, 2013. http://www.catalystsearchmarketing.com/how-rich-snippets-can-improve-your-ctr/. back
  15. Nigel Thrift, Knowing Capitalism, London, 2005, 224 back
  16. This is outlined in Dr. Lop Lop ‘Somebody else's cat: A study in the protohistory of the internet cat meme,’ In Katrina Sluis (ed.) For the LOL of Cats: Felines, Photography and the Web London, expected 2013. back
  17. See, for example the oft cited study by Jacob Nielsen, ‘Participation Inequality: Encouraging More Users to Contribute’ http://www.useit.com/alertbox/participation_inequality.html and Hitwise’s research reported in Bill Tancer, ‘Who’s Really Participating in Web 2.0.’ Time. Apr. 25, 2007 back
  18. Accessed June 3, 2013. http://www.time.com/time/business/article/0,8599,1614751,00.html

  19. Brad Troemel, ‘The Accidental Audience’, The New Inquiry, March 14, 2013. Accessed June 3, 2013. http://thenewinquiry.com/essays/the-accidental-audience/ back
  20. Jodi Dean, Blog Theory: Feedback and Capture in the Circuits of Drive, Cambridge, 2010, 55 back
  21. Lipovetsky, Hypermodern Times, 32 back
  22. Dean, Blog Theory, 36 back
  23. Bauman and Lyon, Liquid Surveillance, 32 back
  24. Hito Steyerl. The Wretched of the Screen. Berlin, 2013, 168 back
  25. Brian Stelter. ‘Upending Anonymity, These Days the Web Unmasks Everyone.’ The New York Times, June 20 2011. Accessed June 2, 2013 back
  26. Sara Reardon, ‘FBI launches $1 billion face recognition project’ New Scientist, Issue 2280 back
  27. This was the experience of my friend, Nic Healy. See: Nic Healy, ‘I got Facejacked!’, C|Net Australia, January 16, 2013. Accessed June 3, 2013. http://m.cnet.com.au/i-got-facejacked-339343023.htm. See also: Douglas Rushkoff, ‘Why I'm quitting Facebook’, CNN International, February 25, 2013. Accessed June 3, 2013. back
  28. http://edition.cnn.com/2013/02/25/opinion/rushkoff-why-im-quitting-facebook/index.html

  29. Benedikt Driessen and Markus Dürmuth, Achieving Anonymity Against Major Face Recognition Algorithms, Cryptology ePrint Archive, 2013. http://eprint.iacr.org/2013/009 back
  30. Sherry Turkle makes this comparison in Alone Together: Why We Expect More from Technology and Less from Each Other. New York, 2011. back
  31. quoted in Holman W. Jenkins, ‘Google and the Search for the Future.’ Wall Street Journal, August 14, 2010. http://online.wsj.com/article/SB10001424052748704901104575423294099527212.html. back
  32. Steyerl, The Wretched of the Screen, 168 back
  33. Finn Brunton and Helen Nissenbaum, ‘Vernacular Resistance to Data Collection and Analysis: A Political Theory of Obfuscation.’ First Monday 16: 5, April 2011. http://firstmonday.org/ojs/index.php/fm/article/view/3493. back
  34. Julian Assange, Jacob Appelbaum, Andy Müller-Maguhn and Jérémie Zimmermann, Cypherpunks: Freedom and the Future of the Internet, New York, 2012, 1-7. back
  35. See for example Hille Koskela, ‘Webcams, TV Shows and Mobile phones: Empowering Exhibitionism’, Surveillance & Society, 2:2-3, 199-215 back
  36. Dean, Blog Theory, 3-4 back
  37. Dean, Blog Theory, 4 back
  38. Geert Lovink, ‘A World Beyond Facebook’, In: Geert Lovink and Miriam Rasch (eds) Unlike Us Reader #8, Amsterdam, 2013, 10 - 12 back
  39. Carlos Barreneche, ‘The Order of Places: Code, Ontology and Visibility in Locative Media’, Journal of Computational Culture, Issue 2, September 2012 back
  40. http://computationalculture.net/article/order_of_places

  41. Barreneche, ‘The Order of Places’ back

Image references

  1. Example back
  2. CV Dazzle, Camouflage from face detection by Adam Harvey back
  3. Anti-drone hoodie by Adam Harvey back