The research committee created by the Dutch Parliament (‘Staatscommissie’) following the child benefits scandal has issued a new report urging the government to stop data-driven profiling for fraud and crime detection, now and in the future.
The Staatscommissie tegen Discriminatie en Racisme is an advisory research committee specifically tasked with identifying initiatives to address discrimination and racism in the Netherlands, explicitly including algorithmic discrimination.
The new report is titled Principles for Profiling (download the PDF) and details the increase in data-driven profiling by the government for (social welfare) fraud detection and policing. The committee lays out clearly, step by step, how these developments are in tension with the rule of law, democratic principles and good governance. All in all, the conclusions are very clear:
Systems for profiling either lead to discrimination or unacceptable risk of discrimination, and they cause a lot of societal damage, as was the case with the child benefits scandal.
The Trouw article connects this report and its clear call to stop the use of profiling to earlier work by the data protection authority, which called out the discriminatory and unlawful use of such systems. Similarly, the scientific council of the police has also advised to halt several types of data-driven policing.
Importantly, the report urges moving beyond such incidents and engaging more fundamentally with the question of whether the use of such data-driven systems is acceptable. A question they answer with an unabated “no”. The risks are simply unacceptable, and alternatives, such as sampling, are available. However, the committee also clearly states that the government is insufficiently open to critique and refuses to look beyond specific cases to the structural dangers connected to these systems.
See: Staatscommissie: ‘Stop met data-gedreven profilering van burgers voor de bestrijding van fraude’ at Trouw.