8: AI & Law enforcement

Question 1

Why is the seperation between national and public security important?

Answer 1

Separation is very important: suppose monopolize information power with enforcement power (mixing information with the ability to use power)
=> same organisation learns a lot about people and would immediately act upon and no one interferes
== wall principle

Ex. Where the wall principle dissapeared: Nazi Germany

National security: threat at the level destroying a country (its democratic order)
-> army
-> intelligence services: goal is to collect intelligence to protect its citizens -> not information on citizens but about external factors

Public security: threat to life of the citizens (Ex. Violent demonstration destroying cars and banks)
-> law enforcement

Public order: even less threat (Ex. Peaceful demonstration)

Question 2

Difference between traditional and predictive policing?

Answer 2

Traditional policing:
* Solving crimes in a reactive manner: at the time or after the crime happens
* less attention to social disorder and the residents quality of life
* requires being on the field, collecting information and evidence
* officer succes measured by number of arrests/responses to calls/ solving crimes
* bureacratic organisational structure: positions of authority/lines of communication
=> succes rate that is used to evaluate determines the daily interaction with people

Predictive policing:
* solving crimes before they happen
* more attention to the past/habits/ any type of disorderly behavior
* requires data/intelligence/ automated analysis
* officer succes also rated in pre-crime interventions
* less supervision, AI-enhanced policing
=> problem: data > evidence: decisions justified by technology and not judicary control

Question 3

Methods of predictive policing?

Answer 3

• Area-based assessment: crimes already committed in that neighbourhood + economic and social fragility of the neighbourhoor + season and time of day = location of future crime
• Individual based risk assessment: economic situation + familiy data + living conditions/habits = who is going to commit a crime

Question 4

Traditional vs predictive policing

Answer 4

Traditional:
* reasonable suspicion: based on specific/objective/articulable facts, must relate to a criminal activity not just the suspect
* judicial control: measures are imposed after approval of a judge
* presumption of innocence

Predictive:
* automated suspicion: based on information of the suspect’s past and has nothing to do with the actions, based on group data, correlation with possible criminal activity rather than a causation with actual criminal activity
* police discretion:
* presumption of guilt

Question 5

Challenges with predictive policing?

Answer 5

• right to non-discrimination and equality:
  -> differentiation grounds that are compromised of a combination of characteristics
  -> profiling techniques often contain of multitude of characteristics which may not be linked to personal traits
  -> complicated correlations (Ex. color of car to traffic violations)
  -> group profiling might not be considered personal, but has a big impact
• right to privacy and personal data protection

Question 6

Why do we need a separate right to personal data protection when we have a right to privacy?

Answer 6

• both rights protect the autonomy and human dignity of individuals and provide them with a personal sphere
• both are an essential prerequisite of freedom for expression/ freedom of peaceful assembly/ freedom of religion
  => differ in formulation and scope:
  1) right to privacy requires non-intervention by states and other actors, while the right to data protection requires active intervention to regulate how to process data
  2) secrecy <=> transparancy
  3) data protection concerns all kinds of personal data, irrespective of the relationship and impact on privacy