Author(s)
Alessandro Acquisti, Ralph Gross and Frederic Stutzman
Source
Journal of Privacy and Confidentiality, Vol. 6, No. 2, pp. 1-20, 2014
Summary
Billions of photographs are posted online, and facial recognition software is becoming more accurate. Researchers used photographs and off-the-shelf software to identify strangers in public places by name and predict their interests.
Policy Relevance
People are unaware that today’s technology allows one to quickly identify strangers.
Main Points
- Google, Facebook, and Apple use facial recognition technology to help users label and sort photographs; the technology is controversial but might be too late to stop.
- Facial recognition software and photos posted to Facebook were used to re-identify users of an online dating site (online to online recognition); about 1 in 10 of the online dating images were identified by their real names.
- Facebook images associated with a certain college were compared with photographs of volunteers taken on campus (online to offline recognition).
- About half wrongly believed that their Facebook images were not publicly available.
- About one third of the volunteers were identified.
- In this study, a phone app was used to compare a photograph of a stranger with a database of social network data, generate a profile that includes her interests, and predict the first digits of her social security number.
- In future, one might search for someone’s face online, just as one searches for a name.
- Successful matches on a large scale require a database containing correctly identified images; social networks provide such a database.
- We instinctively expect that strangers in a crowd will not be able to identify us, but this might not be true in future.