Risks of Profiling and the Limits of Data Protection Law - Discrimination and Privacy in the Information Society - page 135

adjust prices of goods and services on the basis of the profile of the individual.
Charging different prices on the basis of particular characteristics (e.g., race, sex,
or sexual preference) is likely a violation of anti-discrimination legislation.
7.2.5 Inaccuracy
A fifth risk associated with profiling is that profiles might be inaccurate. In par-
ticular there is the problem of 'false positives' and 'false negatives'. This means
that people that in fact do not fit the profile are fitted within it (a false positive), or
people that fit the profile are left outside of it (false negative). False positives and
false negatives occur for various reasons, for instance because insufficient data is
available, or the data is inaccurate. False positives and false negatives are a
particular problem in automated decision making since there is no human inter-
vention and it is not an adversarial process where both sides are heard. This is
troublesome as it places the burden of proof on the side of the data subject: they
must prove that they do or do not fit the profile.
7.2.6 Abuse
A final risk associated with profiling is that data controllers or third parties (for
instance hackers) abuse profiles and/or the information contained therein. Possi-
bilities for abuse arise in particular when the profile can be linked to an identified
individual. A profile could for instance be made public leading to reputational
damage for the data subject (e.g., the data subject is exposed as a dodgy debtor),
or the (personal) data contained in the profile could be used for fraudulent
purposes.
7.3 Privacy and Data Protection in Light of Profiling
To mitigate the risks mentioned in the previous paragraphs, traditionally the right
to (informational) privacy is invoked. The right to informational privacy acts as a
boundary against the free flow of information and thus ensures the protection of
personal information. An important aspect of informational privacy is personal
data protection. In particular in the context of the private sector, data protection
legislation has become the most important aspect of informational privacy protec-
tion. Van den Hoven (2008, p. 311) lists four different moral reasons for protect-
ing personal data. They are: 1) protection against information based-harm, 2) pro-
tection against informational inequality, 3) protection against informational
injustice and, 4) the protection of moral autonomy.
Information based-harm
Because information can be used to cause harm (e.g., identity theft, fraud) or other
serious disadvantages to data subjects, personal data needs to be protected from
access by parties who wish to cause harm using personal data. Data protection sets