of cars passed our sensors, with no results. After about an hour the first alert was
generated. It appeared to be an ambulance of the Maastricht University Hospital,
which had just delivered a patient at the Rotterdam Hospital. The driver was not
amused, nor was the motor-policeman who carried out the selection, while “knowing
perfectly well that this is not what he was looking for”. At the end of that day how-
ever, out of as few as ten vehicles that were stopped as many as six proved to be se-
rious drugs-traffickers. Caught red-handed, each carrying over a kilogram of soft and
hard-drugs. These results convinced even our most sceptical colleagues. As one of
them stated, “It is almost like Christmas day, the presents are delivered and we only
have to unwrap them”.
One week later, at a similar occasion, a taxi was selected based on information
generated by the profile. Under normal circumstances taxis are never stopped. This
time, too, the intuitive response was to let it pass. But it met all criteria of the profile.
Additional information from a cop with local knowledge indicated that despite sec-
ond thoughts the taxi could be worth inspection. It resulted in a find of over 1.7 kilo-
grams of hard-drugs.
After using this profile for a sustained period, results started to decline. The
drug-traffickers started to deviate from their usual routes, thus avoiding the temporal
sensor-network of the police. The most logical route, however, had been compro-
mised, forcing them to take deviant (and a bit awkward) routes. If the police would
be able to deploy sensors on these routes as well, their detection would be easy.
The operations did fuel a healthy public debate about the application of sen-
sor-technology by the police, the mandate of the police to stop a vehicle based on
automated knowledge-rules, and the violation of privacy regulations due to the proc-
essing and alleged storage of large quantities of privacy-related data. Although the
profile-based approach was approved in court, it was concluded that current law did
not provide sufficient clarity for using augmented reality applications in police
Policing science distinguishes various knowledge- and intelligence-disciplines that
are specialized in dealing with various abstraction levels, focal areas, and analysis
techniques (Gottschalk 2008; Holgersson and Gottschalk 2008; Innes et al. 2005;
Ratcliffe 2008). Discussing KBP, profiling, and augmented reality in relation to
these disciplines is beyond the point and scope of this chapter. Instead, we limit
the discussion to the implications of our contribution to the emerging concept of
KBP with respect to the handling of large data-collections in relation to the real-
time discrimination of criminal phenomena, including the impact on privacy.
9.4.1 Databesity: The Ever Present Hunger for Larger Databases
The positivist epistemological foundation of ILP and KBP falls short in acknowl-
edging and dealing with the idiosyncratic, contextual, and dynamic nature of their