structure

1. Three examples
     - Behavioural Targeting
     - Social Scoring
     - Signature Strikes
2. Problem description from data
     protection perspective
3. Your perspective

https://netfog.de/pad/p/isd

Behavioural Targeting

• advertise and offer discounts 
  
• try to meet the interests of the customers
• to change their behaviour by 
  subtle triggers

what target does

pregnancy prediction score is computed based on the 
a correlation of 25 items

which items are bought in which quantities during 
a single shopping tour/by an indentifiable customer

the statistics

item-to-item relations

individual shopping
profiles based on
credit and bonus cards

data sources

Img Source: Wikipedia

1. analyse the data they have 
(as training set) and 

2. match profiles of new customers
with that 

social credit scoring

Business Model

Give micro-credits to people.
height and reates are dependent on the score 
computed from "big data"

data sources

• location data
• behavior data
• social network context

scoring = risk management

• not interested in a specific details but in correlation

from wikipedia: redlining

social scoring is intransparent

While every user can ask for their current SCHUFA score
this is not possible for live computed scores. 

signature strikes

• their social network
  
• their communication behavior
• location data
  
  

potential terrorists
are selected by

Other Examples

- Insurance Companies (car, health)
- security business (passenger control)
- financial service (banks, e-commerce)

questions so far?

• do not care for the individual at all
• personalization is more based on statistics than
  on persons
  
• during the process of data analysis there is no need 
  for a reference to an individual

What those practices
have in common: John|Jan Doe!

informational self-determination?

what has this to do with

privacy

general notions of privacy are bound to physical space  

"The right of the individual to decide what information
about himself should be communicated to others and
under what circumstances."
(Westin 1970)

informational self-determination

"A society would not be compliant with the right to
informational self-determination [...], if citizens would
no longer be able to know who knows what about
them in which occasion."
(German Federal Constitutional Court 1983)

"To protect informational privacy is important for
persons, because it is constitutive for their self-concept
as autonomous person (in it's known borders) to have
control over their grandstanding. To have control over
how they present or produce themselves; how they
undestand themselve; how they want to be understood
in a specific context. And therefore: how to act
in that context."
(Rössler, 2005) 

as basis for liberal societies

data protection principles

1. lawful processing
2. consent
3. stick to purpose
4. data minimization
5. transparency
6. data security
7. control

(Bizer 2007)

But:

In all european laws that refer to informational self-
determination it is bound to an individual (data subject)

Martin Degeling bought bread and a bootle of beer
at konsum in dresden on 02.10.2013 

example: personal information 

XR2201 bought bread and a bootle of beer
at konsum in dresden on 02.10.2013  

example*: anonymous information

* simplified

data protection goals

1. availability
   
2. integrity
3. confidentiality
4. transparency
5. in-concatability
6. intentionability

(Rost 2009)

the information could not be used to send advertisements
to me;
but still influence what advertisements others get to see

even if I never make a credit enquiry;
my job position would still make a difference if my 
friend would do so

"To protect informational privacy is important for 
persons, because it is constitutive for their self-concept 
as autonomous person (in it's known borders) to have 
control over their grandstanding. To have control over 
how they present or produce themselves; how they 
undestand themselve; how they want to be understood
in a specific context. And therefore: how to act 
in that context."
(Rössler, 2005) 

Dividuals

"Dividuals have attributes that are of interest to
the corporations and to the biopolitical police force,
but their identity is not usually of concern
unless a threat is discerned."
(Morris 2012)

 (Deuleuze 1990)

in german law

§6a BDSG (Federal Data Protection Law)

prohibition of automatic decisions
concerning an invididual.

Sources