Twitter is a global social media website best used for publicly disseminating short text messages called tweets. Following recent revelations in the Middle East (a.k.a. the Arab Spring) during which civilians used Twitter and other social media sites to organize massive demonstrations against their governments, Twitter introduced a new policy called “Country-Withheld Content” which enables Twitter to reactively withhold content from users in specific countries while keeping it available in the rest of the world.
During this talk I present analysis of Twitter's censorship mechanism, discrepancies in their published data, and methods for viewing withheld content from inside and outside targeted countries.
I also show evidence of censorship and privacy leaks in tweet structure obtained using Twitter public REST and Streaming APIs. Finally I provide an overview of our ongoing work in developing solutions to identify hidden censored tweets and sensitive users in the Middle East.