SEMANTIC COMPUTING AND PRIVACY: A CASE STUDY USING INFERRED GEO-LOCATION

GERALD FRIEDLAND, JAEYOUNG CHOI
2011 International Journal of Semantic Computing (IJSC)  
This paper presents an experiment that allows the inference over data published in social networks, resulting in a potentially severe privacy leak, more specifically the inference of geo-location resulting in the potential of cybercasing attacks. We present an algorithm that allows the inference of the geo-location of YouTube and Flickr videos based on the tag descriptions. Using the locations, we find people where we can infer both the home address as well as the fact that they are currently
more » ... vacation, which makes them potential targets for burglary. By doing so we repeat an experiment from the literature that was originally meant to show the potential dangers of geo-tagging but replacing the geo-tags with Semantic Computing methods. We conclude that the only way to tackle potential threats like this is for researchers to develop an enhanced notion of privacy for Semantic Computing.
doi:10.1142/s1793351x11001171 fatcat:5dpor5sskfdevouyqjt4ip5kk4