Social Identity Verification

With identity verification a central tech challenge to trustworthy online collective decision-making, the Ethelo Team has been brainstorming ways that social networks networks can effectively achieve identity verification for a centralized decision-making or voting system. This discussion paper is a preliminary overview on how algorithms could analyze social graphs to create a ‘certainty’ factor for online identity.

  1. Allow the decision-making platform to integrate social network information, so that participants can connect with friends and acquaintances they might know on Facebook, Linkedin, Google+, Twitter et cetera. The underlying assumption here is the voting or decision-making process will be something such as Ethelo that uses social networks as part of the decision-making process.
  2. An identity verification algorithm will assign a “confidence score” to each user on the system, by analyzing their social network connections. Trulio is an example of this type of algorithm technology. The network analysis algorithm would look for ‘fake’ profiles by looking for characteristics that distinguish them from a real person’s network presence, such as isolation of nodes. Also, the algorithm will look for ‘key guarantors’, those individuals in a person’s social network whose guarantee of true identity will be the most significant, both due to their place in the network and the confidence the system already has in their identity.
  3. Support but not require “official” verification processes, such as cross-checks with government databases, at the users’ option. For example, a user on the system could submit their name and address, and the government body could cross-check that with official records and post by letter an authorization code to the registered address for that person, which the individual could then enter to be  be confirmed on the system. People who are confirmed through a high-accuracy process would have higher impact as guarantors for others due to the high confidence rating.
  4. Access the capacity of people to verify each other. If a user receives a low confidence rating using the verification algorithm, the platform could look at their social network, and identify those acquaintances that, if verified, will be most influential in raising confidence of the persons’ identity. Then the system would invite one or more of those contacts to verify the user’s identity. Verification would consist of presenting the guarantor with a photo and the name of the person, also perhaps their location of residence, and let the guarantor choose how confidently they will attest to their name and identity. The system can be programmed to manage a guarantor process that is continually looking for ways to increase confidence (or falsify it) of the lowest confidence members. For members whose ID confidence cannot be increase through guarantor verification, other processes can be brought into play (such as credit cards, ID confirmation, release of information forms).

This process could be used to assign an accuracy rating to a voting result, based upon the confidence rating assigned to the participants. If for example, one particular outcome received 10% more support than another outcome, and the system is 98% confident of the identities of all those who participated, then the result could be deemed valid. While not a “perfect” system, it could be used in conjunction with other security systems to add another layer of confidence to results.

This process has the advantage of easy scalability without a time-consuming, government-based process. Moreover, it will grow stronger and more accurate the more people are added to the system, and the more alternative ways of verifying identity are added.

If you have feedback or are interested in this project, please contact

2018-03-20T02:05:02+00:00May 17th, 2012|General|0 Comments

Leave A Comment