User-centred privacy inference detection for smart home devices

Alexia Dini Kounoudes, Georgia M. Kapitsaki, Ioannis Katakis, Marios Milis

Abstract: In the smart home, vast amounts of data are being collected via various interconnected devices. Although this assists in improving the quality of life at home, often the user is not aware of the details concerning data collection apart from the information available on the provider privacy policy. It is however important to put the user inside this loop of information, so that she is well informed on possible uses of the data and the potential risks that this may entail. Previous works have identified user activity inside the smart home and have pointed out privacy threats. In this work, we go one step further by offering data inference techniques and giving this information back to the user. We use a number of machine learning techniques to draw conclusions about the user routines or activities and we inform the user about our findings concerning data inferences through a dedicated web application. Our aim is toward user-centred privacy and is a proof of concept that can be reused by smart home and Internet of Things service providers in general in order to improve the services offered to the end-users. Our results indicate that a large number of data inferences are possible by using a combination of techniques.