Algorithmic Identification : Bookmarks and Identicon..?

• Upturn’s Automation & The Quantified Society
• The odd reality of life under China’s all-seeing credit score system
• How the QR code conquered China

After reading these articles, I kept asking myself

“What does it mean to use personal data to create a unique representative mapping of individuals?”

If we think about China’s all-seeing social credit score system, we could abstract the system down to a simple function that takes a sequence of human behaviors and outputs a number after going through its internal algorithm.

For example,

Input :
{behavior:"cheated in game",time:"09/10/18"},
{behavior:"harmed two neighbors",time:"09/10/18"},
{behavior:"earned $100",time:"09/10/18"},
{behavior:"protested against government",time:"09/10/18"},
{behavior:"destroyed public property",time:"09/10/18"},

Output : 123

So this is a mapping from a sequence of behaviors to a credit score. In the end, we are left with no details but with an abstraction of data that can’t be reversed back to its original form. (X -> Y is possible, but Y->X is impossible even if we know the internal algorithm!)

The QR code is also a type of abstraction and a mapping of personal information (account number, or ID number) to a representation (QR Code). It’s interesting that QR code serves as a means of identification that people feel more comfortable sharing with strangers because the private personal data is encapsulated in the representation.

I feel motivated to dig deeper into this notion of creating identification from personal data and the subtle implications behind this idea.

Bookmarks (for internet browsing) interest me because they form a unique collection of an individual’s interests, history, and life style. I think of them as our web finger prints that might potentially serve as an identification when abstracted. They are even more interesting because they are super private. (So the main challenge would be.. how do I collect people’s bookmarks?) So I did a little bit of hunch hunting.

I arrived at a potential idea of converting raw bookmarks texts into Identicons, just like the default profile image we see on services like Github and Wordpress. Below are some links to websites that I found helpful brainstorming ideas for bookmarks and Identicons.

• https://blog.github.com/2013-08-14-identicons/
• https://barro.github.io/2018/02/avatars-identicons-and-hash-visualization/
• https://www.npmjs.com/package/identicon-github
• https://blog.byndyusoft.com/generative-avatars-16bdb15f872d
• https://en.wikipedia.org/wiki/Identicon
• https://github.com/donpark/identicon
• https://blog.docuverse.com/tag/identicon/
• https://haacked.com/archive/2007/01/22/Identicons_as_Visual_Fingerprints.aspx/
• https://github.com/stewartlord/identicon.js/tree/master

While I’m left with a lot more exploration to further dive into this topic, I’m looking forward to transformation of bookmarks to Identicons!

I’m looking for teammates and please shoot me an email (slyu@olin.edu) if you are interested in this idea!

Further things to think about :

• What algorithm am I going to use as the “hash” function?
• What form of bookmarks data am I going to use? (Favicon images, bookmarks html export, raw texts, titles)
• What context and form should I provide to the audience?
• What style makes sense?
• How am I going to gather bookmark data? (Survey? Ask friends? Helpme email?)