Paper Title: Privacy Risks Emerging from the Adoption of Innocuous Wearable Sensors in the Mobile Environment
Authors: Andrew Raij, Animikh Ghosh, Santosh Kumar and Mani Srivastava
Authors Bios:
Andrew Raij is an Assistant Professor in the Department of Electrical Engineering at the University of South Florida. Previously, he held post-doctoral appointments at the University of Memphis and the University of Florida. He received a PhD in Computer Engineering from the University of Florida and an M.S. and B.S. in Computer Science from the University of North Carolina at Chapel Hill and Northwestern University resp.
Animikh Ghosh is a Junior Research Associate at SETLabs for Infosys Technologies Ltd. He is interested in wireless sensor networking, privacy risks involved in participatory sensing and database designing.
Santosh Kumar is an Associate Professor at the Department of Computer Science at the University of Memphis. He lead the Wireless Sensors and Mobile Ad Hoc Networks Lab. He is engaged in both theoretical and systems research. In theoretical, he is recognized for his work on coverage and connectivity.
Mani Srivastava is on the faculty at UCLA as Professor in the Electrical Engineering Departmnt, with a joint appointment as Professor in the Computer Science Department. Before joining UCLA in 1997, he worked for about four and a half years at the Networked Computing Research Department at Bell Labs in Murray Hill, NJ.
Presentation Venue: CHI '11 Proceedings of the 2011 annual conference on Human factors in computing systems that took place at New York (ACM)
Summary:
Hypothesis: In this paper, the authors present the concept of wearable sensors that are revolutionizing healthcare and science by enabling capture of physiological, psychological, and behavioral measurements in natural environments. They mention about their study where they conducted a study to assess how concerned people are about disclosure of a variety of behaviors and contexts that are embedded in wearable sensor data. The authors analyze the data from three perspectives:
1) They assess how disclosure of different behaviors and contexts affect participant concern levels as their stake is increased in the data
2) They evaluate the impact of applying various restrictions and abstractions on concern level.
3) They assess the impact of re-identification on the concern level as the role of the data consumer is varied from the research team to the general public
How the hypothesis was tested: 66 participants were recruited from the student population at a 20,000+ student university in the United States. The participants volunteered to join one of two groups, one with no personal stake in the data (NS) or the one with a personal stake (S) in the data.
For three days, Group S collected physiological, behavioral, and psychological data using the AutoSense sensor system as they went about their normal everyday life. At the end of the 3 day period, group S participants completed a privacy questionnaire assessing their concern regarding disclosure of selected behaviors and contexts with various restrictions and abstractions applied.
Group NS participants had no exposure to continuous physiological, behavioral and psychological data collection. They did not wear AutoSense and did not review any data collected by it. They only completed the same privacy questionnaire as Group S. This allowed a between-subjects comparison of concern levels between participants with no personal stake in the data, Group NS, and participants with a personal stake in the data, Group S.
Results: The authors analyzed participant data with respect to the three goals of the study as discussed in the preceding section. Two-taied t-tests were used to test for significant differences (p<0.05). In comparisons between Groups NS and S-Pre (different populations), unequal variances were assumed. In comparisons between Groups S-Pre and S-Post (same population), a paired t-test was used.
The results revealed that participants were most concerned about sharing physical and temporal context together. There was a trend of increasing concern for place and timestamp, with the lowest concern for just sharing the behavior or context, the next level of concern for reporting the place or time of the behavior or context, and the highest level of concern for reporting both the place and time of the behavior or context.
Discussion:
Effectiveness: The results of this experiment indicate that people cannot understand the potential threats in the data unless they have a personal stake in it. The experiment performed by the authors provides a very good idea to us that the community should examine disclosure of physiological context more closely. The paper suggests that the community should also examine how data consumers perceive privacy issues and what aspects of the data make it useful. Such a study would provide a better understanding of how to tradeoff behavior privacy and utility for physiological, psychological, and behavioral data collected by personal sensors.
Reasons for being Interesting: The study performed by the authors in this paper is imperative considering the amount of data being shared in the technologically advanced generation today.
Faults: The paper has many limitations but I did not really find any faults about it.
Authors: Andrew Raij, Animikh Ghosh, Santosh Kumar and Mani Srivastava
Authors Bios:
Andrew Raij is an Assistant Professor in the Department of Electrical Engineering at the University of South Florida. Previously, he held post-doctoral appointments at the University of Memphis and the University of Florida. He received a PhD in Computer Engineering from the University of Florida and an M.S. and B.S. in Computer Science from the University of North Carolina at Chapel Hill and Northwestern University resp.Animikh Ghosh is a Junior Research Associate at SETLabs for Infosys Technologies Ltd. He is interested in wireless sensor networking, privacy risks involved in participatory sensing and database designing.
Santosh Kumar is an Associate Professor at the Department of Computer Science at the University of Memphis. He lead the Wireless Sensors and Mobile Ad Hoc Networks Lab. He is engaged in both theoretical and systems research. In theoretical, he is recognized for his work on coverage and connectivity.
Mani Srivastava is on the faculty at UCLA as Professor in the Electrical Engineering Departmnt, with a joint appointment as Professor in the Computer Science Department. Before joining UCLA in 1997, he worked for about four and a half years at the Networked Computing Research Department at Bell Labs in Murray Hill, NJ.
Presentation Venue: CHI '11 Proceedings of the 2011 annual conference on Human factors in computing systems that took place at New York (ACM)
Summary:
Hypothesis: In this paper, the authors present the concept of wearable sensors that are revolutionizing healthcare and science by enabling capture of physiological, psychological, and behavioral measurements in natural environments. They mention about their study where they conducted a study to assess how concerned people are about disclosure of a variety of behaviors and contexts that are embedded in wearable sensor data. The authors analyze the data from three perspectives:
1) They assess how disclosure of different behaviors and contexts affect participant concern levels as their stake is increased in the data
2) They evaluate the impact of applying various restrictions and abstractions on concern level.
3) They assess the impact of re-identification on the concern level as the role of the data consumer is varied from the research team to the general public
How the hypothesis was tested: 66 participants were recruited from the student population at a 20,000+ student university in the United States. The participants volunteered to join one of two groups, one with no personal stake in the data (NS) or the one with a personal stake (S) in the data.
For three days, Group S collected physiological, behavioral, and psychological data using the AutoSense sensor system as they went about their normal everyday life. At the end of the 3 day period, group S participants completed a privacy questionnaire assessing their concern regarding disclosure of selected behaviors and contexts with various restrictions and abstractions applied.
Group NS participants had no exposure to continuous physiological, behavioral and psychological data collection. They did not wear AutoSense and did not review any data collected by it. They only completed the same privacy questionnaire as Group S. This allowed a between-subjects comparison of concern levels between participants with no personal stake in the data, Group NS, and participants with a personal stake in the data, Group S.
Results: The authors analyzed participant data with respect to the three goals of the study as discussed in the preceding section. Two-taied t-tests were used to test for significant differences (p<0.05). In comparisons between Groups NS and S-Pre (different populations), unequal variances were assumed. In comparisons between Groups S-Pre and S-Post (same population), a paired t-test was used.
The results revealed that participants were most concerned about sharing physical and temporal context together. There was a trend of increasing concern for place and timestamp, with the lowest concern for just sharing the behavior or context, the next level of concern for reporting the place or time of the behavior or context, and the highest level of concern for reporting both the place and time of the behavior or context.
Discussion:
Effectiveness: The results of this experiment indicate that people cannot understand the potential threats in the data unless they have a personal stake in it. The experiment performed by the authors provides a very good idea to us that the community should examine disclosure of physiological context more closely. The paper suggests that the community should also examine how data consumers perceive privacy issues and what aspects of the data make it useful. Such a study would provide a better understanding of how to tradeoff behavior privacy and utility for physiological, psychological, and behavioral data collected by personal sensors.
Reasons for being Interesting: The study performed by the authors in this paper is imperative considering the amount of data being shared in the technologically advanced generation today.
Faults: The paper has many limitations but I did not really find any faults about it.
No comments:
Post a Comment