Smartphones have created a world in which seemingly everything is recorded, but new “lifelogging” wearable cameras will make even fewer moments pass unseen. These devices, such as Google Glass, Narrative Clip and Autographer, are always on, and always recording what is happening around the wearer. These cameras raise important privacy concerns, not just for the wearer, but also for bystanders who happen to be in the background of photos or videos.
To help understand the privacy implications of these new devices, and to eventually help people protect their privacy as these devices become popular, School of Informatics and Computing assistant professors Apu Kapadia and David Crandall along with Indiana University researchers Roberto Hoyle, Robert Templeman, and Steven Armes, and Dartmouth College’s Denise Anthony conducted a study of people’s lifelogging behaviors and perceptions. The study focused on how and why lifeloggers share certain photos, what kinds of private content is captured in lifelogging, how people manage their personal privacy, and how bystanders perceive such devices.
For the study, a total of 36 students were asked to wear a lifelogging camera, which took a picture every five minutes throughout the day for a week. Users had several ways of protecting their privacy. They could choose to pause collection for a few minutes, they could delete the last few minutes of photos, or they could take the camera off or physically block the lens.
Later they reviewed the images on a website and answered questions about them. They could delete photos that were highly private or were blurry, and were also asked to mark which photos they would and would not share with different types of people (like close friends or the general public). The study collected a total of almost 15,000 images from the 36 participants.
The results showed that most participants were willing to share the bulk of their images. The vast majority of the photos were banal, everyday moments and participants saw no good reason not to share them. But several photos with private information were also captured, including computer screens showing emails and grades, and participants were unwilling to share many such photos. Most participants were clearly aware of privacy risks, hitting the pause button an average of 6 times per day, and also covered the camera lens regularly. Using the restroom or being in another place where recording was prohibited were the two main reasons, but data collection was also paused because there was someone nearby or there was nudity nearby. In other words, people showed concern for the privacy of other people.
Bystanders, meanwhile, seemed surprisingly tolerant of the lifelogging cameras. One college class debated the use of the camera and voted to allow it, while two other participants asked for and received permission from their sororities to record their activities.
The results of the study suggest that although lifelogging could provide an unfiltered view of the day-to-day world around us, users of the devices are conscious of what they are recording and whom they might be impacting. Through a recent $1.2 million grant from the National Science Foundation, the IU and Dartmouth team plans to build on the results of the study to help design new ways to help people manage their privacy in the era of wearable cameras. For instance, they are exploring technical means of using computer vision technology to detect and censor private content in images. Their overall goal is to help understand and address wearable camera privacy before these devices become widely adopted.