D7.4: Implications of profiling practices on democracy

Title: JUST WHAT IS IT THAT MAKES TODAY’S PROFILING SO DIFFERENT, SO REPELLING?

Just what is it that makes today’s profiling so different, so repelling?

In 1956, Richard Hamilton made a photo-collage profile of a modern home and gave it, with a mere touch of irony, the title: ‘Just What Is It That Makes Today’s Homes so Different, so Appealing?’ The home has changed forever, but why does it look so cool – is it the television, the vacuum cleaner, the Ford lampshade, or is it the pin-up on the couch or the semi-nude bodybuilder who seduces us to play pop?

When you read the texts of Hildebrandt and Gutwirth & De Hert on profiling and the rule of law, the message is crying out at you that today’s profiling is different, and that it is repelling because it threatens the rule of law. But just what is it that makes today’s profiling so different, so repelling? Is it the large-scale collection of personal – and non-personal – data in data warehouses, is it the data mining that seeks to uncover relations that are imperceptible to the human mind, is it the marketing of the resulting profiles, or is it the use of these profiles in various kinds of situations? Or is it perhaps the impact of the use of profiles on the human person’s sense of self, her ipse-identity?

We should be careful in making distinctions, both conceptual and practical, before we can begin to denounce profiling as a threat to the rule of law. We could start with three stages of profiling: 

1. pre-profiling: the collection and storage of data; 

2. profile-making: the analysis of data collections in order to make profiles; 

3. profile use: the application of a profile in a concrete case. 

However, to be able to assess the risk of profiling on the rule of law, we need more subtle distinctions. For instance: 

1. pre-profiling: the collection and storage of  

   1. non-personal data, 

   2. personal data; 

2. profile-making: the analysis of data collections in order to make: 

   1. group profiles, 

   2. personalised profiles; 

3. profile use: the application of a profile to: 

   1. make a general rule: 

      1. in science, 

      2. as a commercial, and marketable, asset, 

      3. that is used by the organisation to make general organisational decisions, 

      4. that is used by the organisation to make specific organisational decisions; 

   2. apply in a concrete case: 

      1. to decide to offer something to a group or not; 

         1. which no-one is really remotely interested in, 

         2. which most of the group members would actually like to have; 

      2. to decide to offer something to an individual: 

         1. at all, 

         2. with a discount, 

         3. with a surcharge; 

      3. to decide whether to grant a request by an individual: 

         1. which is not vital to the individual, 

         2. which is vital to the individual. 

(Note the distinction in type 3 between a) a profile-based rule that is used for decision-making, and b) the profile itself that is used for decision-making. The use of rules is not evidently included in the term profiling, since the profile has done its work and is discarded, but from the point of view of law or legal principles, it may be equally relevant.)

This is just a sample taxonomy, and a fairly simple one at that, but it shows the importance of making distinctions in order to know what we are really talking about. I think, for instance, that the distinction under 3bii is relevant, because it makes a difference whether a profile is used to make an offer to someone or not (the profiled is not put in a position to choose), whether the offer is made with a discount (the profiled is rewarded for not fitting the profile), or whether the offer is made with a surcharge (the profiled is punished for fitting the profile). And making group profiles to generate general rules for scientific purposes is something really different from applying an individual profile to deny someone a request. There is nothing wrong with many of the profiling actions in this taxonomy; only certain types of profiling are noteworthy because they may affect people’s lives or fundamental legal principles. So, the first step is to be precise in indicating what type of profiling we are concerned with. 

The next step is to argue why a certain type indeed impacts on people’s lives. The previous texts argue rather lightly that profiling has serious consequences for human beings, but is that really the case? In my view, consequences occur only in profile use, and serious consequences probably occur only in type 3biii2: applying a profile when deciding a request that is vital to someone. I shall not argue that this does not happen, nor that ICT and the resultant correlatability of data has not greatly facilitated this type of profile use. However, if it is to be shown that profiling poses significant threats that differ radically, rather than gradually, from the age-old application of profiling in the past, we need convincing qualitative examples and quantitative data of serious negative effects of profiling. I have found neither in the previous texts. Here, then, is a clear goal for further research: map the actual consequences of profiling on people’s lives in real life.