Against self-fulfilling prophecies 
Contemporary mechanism of dividualization in
cybernetic capitalism and what to do about it
Martin Degeling - Ruhr University Bochum - 05.12.2013
  • Profiling and Tracking mechanisms are a threat to decisional
    and informational privacy

  • Working with aggegated and anonymized data make it difficult to critique
    from a indiviual-centric position based on liberal hypotheses

  • The cybernetic approach of those technologies does not care for individuals
    and sees them as systems that can be regulated like any other

  • My study should 





  • The tools and theory are work-in-progress and any comment, suggestion or
    critique is very welcome
  1. help to shed light on these mechanisms
  2. Demonstrate the inappropriateness of the cybernetic hypothesis
  3. Offer a collective way to obfuscate individual behaviour and disturb
    profiling
On One Slide
Predictive Policing
Social-Scoring, Behaviour
based insurances, Obamas Election Campaign ..

Other examples
Example: Behavioural Ad-Targeting
"[Mr. Hughes is a] frequent shopper, [...] he logs on to Facebook and sees that his
friend Ella has just become a fan of Bryce Computers, an imaginary electronics
retailer and Acxiom client. Ella’s update prompts Mr. Hughes to check out Bryce’s fan
page and do some digital window-shopping for a fast inkjet printer.

When Mr. Hughes follows a link to Bryce’s retail site [...] the system recognizes him from
 his Facebook activity and shows him a printer to match his interest. He registers on
the site, but doesn’t buy the printer right away, so the system tracks him online.
Loe and behold, the next morning, while he scans baseball news on ESPN.com,
an ad for the printer pops up again.

That evening, he returns to the Bryce site where, the presentation says, “he is instantly
recognized” as having registered. It then offers a sweeter deal: a $10 rebate and free
shipping. [...]

Correctly typecast, Mr. Hughes buys the printer."
Taken from: Singer, N.: Acxiom, the Quiet Giant of Consumer Database Marketing, New York Times , (2012). Last seen online: 28.11.2013

Implications of the examples
  • use of measurable attributes of a person
  • set them in relation to statistical information
    about groups that share those (or similar)
    attributes, to conclude additional information
  • offer products or information tailored to the
    assumed interest of that combined profile

Possible negative effects
Price discrimination/redlining




Filter Bubble



Prejudices/self-fulfilling prophecies
Danna, A., Gandy, O.H.: All That Glitters is Not Gold: Digging 
Beneath the Surface of Data Mining. Journal of 
Business Ethics. 40, 373–386 (2002).
 Sweeney, L.: Discrimination in Online Ad Delivery. Social
Science Research Network, Rochester, NY (2013).
Pariser, E.: Filter Bubble: Wie wir im Internet entmündigt 
werden. Carl Hanser Verlag GmbH & Co. KG (2012).
img src: wikipedia
Cybernetic dividuals..
Cybernetics
Re-read today² as a founding area of modern capitalism
Scientists from various areas started in the 50s to  
"[...] develop a theory of regulation, governance and control" ¹

1. Pias, C.: Die kybernetische Illusion. In: Liebrand, C. and Schneider, I. (eds.) Medien in Medien. DuMont, Köln (2002).
Everything can be understand as a black boxes that can be regulated by
controlling input and using the output as feedback.
Norbert Wiener (img src)
2. Die technologische Bedingung: Beiträge zur Beschreibung der technischen Welt. Suhrkamp, Berlin (2011).
Developers of technology like (behavioural) targeting, profiling or
scoring assume that a measurable number of attributes
of individuals
 allow conclusions about future behaviour and try to
influence the way one is able to interact with 'the world'
based on that estimations
'Dividuals' describes the fragmentation of individuals
to attribute based sub-entities that are
exposed to power relations¹
1.  Deleuze, G.: Postskriptum über die Kontrollgesellschaften. L’autre journal. 1 (1990)
2. Esposito, E.: Die Fiktion der wahrscheinlichen Realität. Suhrkamp, Frankfurt am Main (2007).

Probabilities form an alternative reality that
feeds back into the real reality² 
Why is this a privacy problem?
"We are not interested in the user, 
but in it's statistical model."¹
Stephan Noller; CEO nugg.ad
 Biermann, K.: Behavioral Targeting: Wie vorhersagbar unser Verhalten ist, Zeit.de, (2010). Translation: MD. Last access: 28.11.2013

It's easy: no person related data, no privacy issue.. 
informational privacy
decisional privacy
locational privacy
1. Rössler, B.: The value of privacy. Polity, Cambridge, UK ; Malden, MA (2005).

privacy is a necessary prerequisite to become (and stay)
an autonomous individual in liberal societies
Rössler's differentiation of privacy
Informational privacy
"The core of the question is about who knows what about a person,
therefore the question is about control over information, that relate to someone; 
control in the sense of regulating access but also to at least beeing able 
to estimate [...], what people or organisation [...] know about someone and 
that s_he can enact sanctions or critique based on that assumptions and expectations"

1. Rössler, B.: "Der Wert des Privaten". In: Privat! pp. 15–32. dpunkt (2003). Translation by MD

This is similar what is described
when the right to informational
self-determination is explained
in Germany

img src: welt.de
[..] freedom rights guaranty (in their know, liberal-democratic
borders) to choose a lifestyle, the decision for a partner and
the right to move and behave, according to ones own will;
but this requires a social space in which one can live
unmolested[..] (Rössler 2001, p 147)
Decisional Privacy
Rössler, B.: Der Wert des Privaten. Suhrkamp, Frankfurt (2001).  Translation by MD
[..]if self-determination, the autonomy of a person, can be understood in a
way, that a person has (in the known borders) the right to 
be the author
of her own biography, then this means in social contexts, that the own life
is not commented or interpreted or even influenced by those that she does
not want to grant the privilege to do so
. (Rössler 2011, p 153)
-> With regard to the described technical phenomena this should be extended
to contexts of interaction between single persons and automatic systems.
  • there are much more constraints to autonomy than privacy 
  • autonomy has to be seen in (social) contexts  and relation
  • the concept of a rational (male) autonomous individual has to be questioned 


nevertheless the concept describes a concept that is the baseline for
discussion about power relations in our societies.
Baseline: Autonomous individuals
Privacy is a prerequisite for  western, liberal and
democratic societies  to allow individuals to
be autonomous individuals
But: The cybernetic hypothesis¹
relies on another concept
1. Tiqqun: Kybernetik und Revolte. Diaphanes, Zürich; Berlin (2007).

  • Where individuals are referenced as dividuals
  • Everything is a system that can be regulated
  • Behaviour can be triggerd by manipulating the input
2. Develop a PET that targets cybernetic elements
  • have a closer look at the black box
  • obfuscate dividuals instead of supporting individuals
Conclusion (for me)
1. Identify the cybernetic assumptions on which
     these technologies are based

What I want to do
Ad-Services as black-box to
  • start by looking at
    the output (the ads)
  • Influence the input to obfuscate
    information that could be used
    for profiling



Online Advertising, because
  • It's making the most marketing for
    it's abilities 
               in profiling
               in influencing behaviour
 
  • It's easy accessible
My Case Study
Analyse Online Targeted Advertising
and obfuscate profiling
TrackTrack
A framwork for evaluating effectiveness
of Anti-Profiling tools
Data Sources for Assumend User Profiles
  • Reddit Profiles based on their API
  • Quantcast and Alexa Statistics about average users
  • Google Interest Categories
img src: quantcast.com
An intervention against profiling of online behavioural ad-targeting
by collective obfuscation

TrickTrack
A combination of Google Sharing and TrackMeNot
Mix profiles to hide in them. But not hide in pure Noise
but in other profiles

  • Exchange identifiers (Cookies)
    to mix profiles

  • Randomly add pages to a users
    history to neutralize an profiled attribute
Status:
  • Working Crawler for Profiles
  • Visualization of results
  • Ability to tag advertisements
Initial Results (from Track Track)
  • Most advertisements are related to 
    general campaigns and not to profiles

  • Those that seem targeted are
    related to products that are
    within the users URL-list

  • There is a lot of 'spam' advertisement

  • Google is omnipresent (>50% of pages) and has assumes
    'interest' very quickly
Targeting
Is this related to the profiles that do not include a lot of shopping pages or to
my setup? Or due to the lack of interaction with the page?

Data Quality
Is it even possible to mix the data of alexa, quantcast and google? Without
e.g. weighting them 

User Transparency
Is it of any use for lay users? e.g. allow importing own history list

Question the Aim
Maybe TrackTrack would only help 'targeters' to improve their services

tbd
On One Slide
  • Profiling and Tracking mechanisms are a threat to decisional
    and informational privacy

  • Working with aggegated and anonymized data make it difficult to critique
    from a indiviual-centric position based on the liberal hypothese

  • The cybernetic approach of those technologies does not care for individuals
    and sees them as systems that can be regulated like any other

  • My study should 





  • The tools and theory are work-in-progress and any comment, suggestion or
    critique is very welcome
  1. help to shed light on these mechanisms
  2. Demonstrate the inappropriateness of the cybernetic hypothesis
  3. Offer a collective way to obfuscate individual behaviour and disturb
    profiling