Who’s That? Don’t Worry–Just Ask Facebook’s Metadata

It was November 1st, and I was hurriedly uploading to Facebook my photos from the Halloween party which I had attended the night before.  As usual, its facial recognition software correctly identified each friend, regardless of their Captain America suits or ice-blue Elsa inspired garb.  Unconsciously, I accepted the service’s suggestions; unconsciously, that is, until I reached an image which gave me pause.  In it, I am standing in between two of my hall mates who not only share the same outfit but also face–they are identical twins.  Although I see with them in person on a daily basis, I rely on their personality quirks or fashion choices to discern with whom I am interacting as to me, they are physically indistinguishable.  Yet, in an instant, Facebook had assigned each of them a name, to which I nervously agreed.  Breathing a sigh of relief when one of them left a confirming comment, I was left scratching my head about how it all worked.  Was it really possible that a website could know my friends better than I?

When it comes to their looks, the answer is almost.  Facebook’s recognition software works by automatically generating metadata on individuals’ faces–which become, to use Anne Gilliland’s term, information objects–each time that they are tagged, taking precise measurements of eyes, lips, noses, and every thing in between at various angles.  According to an article published in 2015 by Fortune magazine, this software can determine users’ identities with 83 percent accuracy.  Humans can correctly identify faces 98 percent of the time.  But this technology is improving as it collects more metadata about each face.  ExtremeTech reports that Facebook has developed a new software, DeepFace, which matches humans’ accuracy levels.

While Gilliland focused primarily on metadata generated by the creator of information objects, such as headings or tags, this is an example of automatic, interpreted metadata generation based on algorithmic analysis of the object.  Despite the fact that the origin of DeepFace’s metadata is inhuman, it functions in a way which is ultimately the same as more traditional information resources.  For example, the measurements recorded by the program on a given photograph come together to reflect the image’s content, or what it contains.  It also performs similar functions, facilitating search and retrieval of images depicting a particular person.  However, the surprising accuracy of DeepFace reveals that continuing to shift metadata formulation from humans to computers will increase its precision, thus making it more useful.



This entry was posted in Week 5 blog post. Bookmark the permalink.

1 Response to Who’s That? Don’t Worry–Just Ask Facebook’s Metadata

  1. Annie Kingman says:


    At first I was confused by how metadata accrues, but your post is a perfect example. It is strange to me that something that seems so static, like data, is a growing and changing entity. These are learning softwares, which means they are becoming more and more powerful. Each additional piece of information adjusts the processor. The question is, like you said, when does it stop? Just how powerful will these programs get, and will they surpass the human mind?

    Your example with the identical twins is perfect. If Facebook had never seen these two faces before, it may have difficulty determining who is who. But, because the data has built upon-itself from previous images, it now has the ability to accurately differentiate between two nearly-identical faces. I cannot fathom how complex this data must be. Out of Facebook’s millions of users, it has categorized the minute facial features of an individual.

    I think there is huge potential with these learning softwares. For example, medical research and development. These data programs can record and compare detail with impossible specificity, and as they learn they can have a positive impact on the next patient.

    But, are there also dangers? Certainly. Facebook’s software can identify us in a split-second, and it has stored this info somewhere. Who has access to this information? If it is stolen and manipulated, there is major cause for concern. It might lead to identity theft, misrepresentation, unauthorized tracking, etc. It will be interesting to see how these kind of programs develop in the upcoming decades.

Leave a Reply

Your email address will not be published. Required fields are marked *