Image
Cybercrime file image
- Credit:
Rendering: www.elbpresse.de / Christoph Scholz /
Flickr / Wikimedia Commons
- License:
CC-BY-SA
Friday, 29 April 2016 - 08:23
Data mining risks profiling ordinary citizens as criminals says Dutch council
Analyzing data gained from data mining can help spot fraud and solve crimes, but comes at the risk of profiling citizens like criminals, according to the Dutch scientific council for government policy WRR in its latest advice to the government, NOS reports.
The government is already analyzing citizens' data - the tax authorities use "big data analysis" to catch fraud. "And we expect the intelligence services do it too, thou we don't know it", Erik Schrijvers of the WRR said to the broadcaster. The council expects the use of big data will only increase, while the government's rules and regulations are still inadequate and regulators have little expertise.
Big data analysis risks citizens being profiled and unfairly placed in certain corners because they exhibit certain characteristics. Personal data that have little effect when stored - like purchase history, call history and travel information - can have implications for citizens' privacy when analyzed.
It can also lead to a "Big Brother effect" on citizens - people start acting differently because they feel watched. "When people feel that they are being permanently observed, it could have a chilling effect on free society", WWR member and former Justice Minister Ernst Hirsch Ballin said.
The WRR is therefore calling on the government to establish additional rules for the use of citizens' private data. The current legislation focuses on collecting data, but the analysis of that data should also be regulated. For example, analysis from an algorithm should not be considered proof in court. And decisions should not be made by algorithms. "Computer says no, or maybe yes, is undesirable", Hirsch Ballin said. "If I get a wrong recommendation for a book or music, then that might be a wasted effort. But if people are arrested or stopped based on algorithms, that goes too far."
The council also suggests giving supervising authorities, such as the CTIVD which supervises the intelligence services and the Personal Data Authority, more capabilities and more expertise.