Can Privacy-Aware Lifelogs Alter Our Memories?

Abstract

The abundance of automatically-triggered lifelogging cameras is a privacy threat to bystanders. Countering this by deleting photos limits relevant memory cues and the informative content of lifelogs. An alternative is to obfuscate bystanders, but it is not clear how this impacts the lifelogger's recall of memories. We report on a study in which we compare viewing 1) unaltered photos, 2) photos with blurred people, and 3) a subset of the photos after deleting private ones, on memory recall. Findings show that obfuscated content helps users recall a lot of content, but it also results in recalling less accurate details, which can sometimes mislead the user. Our work informs the design of privacyaware lifelogging systems that maximizes recall and steers discussion about ubiquitous technologies that could alter human memories.

DeepFakes for Privacy: Investigating the Effectiveness of State-of-the-Art Privacy-Enhancing Face Obfuscation Methods

“…An issue with inpainting however is that it removes information that individuals were present in the original photo. This may impact the users' memories in a negative and unethical way [6]. There are also some situations in which inpainting may not be possible.…”

“…A study by Elagroudy et al [6] investigated the effect of privacyaware obfuscations on user's memories. They found that ambiguous life-logs with obfuscations may distort memories.…”

Impact of Privacy Protection Methods of Lifelogs on Remembered Memories

“…The experiment was repeated twice in which participants first took part in an eventful interaction session and then returned after 4-5 days to recall memories when viewing (1) 20 unaltered photos, (2) obfuscated versions of the 20 photos where persons are blurred, and (3) five out of the 20 original photos after deleting private ones. Our experiments are largely inspired by the methodology of Elagroudy et al 's work [15] and Li et al 's work [37]. Our work complements [37] by validating the user preferences for the obfuscation methods in the context of lifelogs as opposed to generic photo sharing.…”

“…Like recent prior work (e.g. [12,15,33]), we investigate the impact within the context of environmental lifelogging, where wall-mounted cameras capture third-person views of the lifeloggers and bystanders. We chose this context for increased relevance as it mimics surveillance scenarios.…”

“…Additionally, photos from such infrastructure cameras are clearer than body-worn lifelogs and have higher utility [11], thus better suited for our target. Due to the exposing nature of lifelogs capturing sensitive information, the privacy of the bystanders and lifeloggers is commonly protected in literature and industry using either blurring or deletion of the sensitive content [15,27,36,37,54]. Thus, our work investigates the impact of both methods on skewing memories.…”