In 2016 we scrapped the public online company registers that tax havens keep as an empty gesture towards transparency. These records (half a million companies and naturally full of gaps) were compiled into a single database in an attempt to identify patterns and links.
The gaps in the database are filled in using algorithms trained to speculate on the forms of the companies. A web scraper identifies keywords and attempts to find images to represent the companies; a neural network trained on scraped business objectives writes a mission statement for each nondescript entity; a Twitter Bot (@OffshoreAZ) repeatedly list these speculative companies, linking to the centralized database as a permanent record of their possible activities: the Offshore A—Z [http://offshoreaz.company].
Following the leaks of the Panama Papers, this database has been cross referenced with the released documents highlighting companies of interest in both datasets.
For the Web Residency we propose to setup a hotline offering callers information on this subset of companies. Greeted by the warmth of Caribbean sounds, the caller will hear the leaked company details: name, country, registration date, mission statement, and so on.
The residency will be an opportunity to increase the fidelity of this repository of financial misconduct: we will train a stronger neural network to provide objectives and other company information based on the increased awareness the Panama Papers provide. This will feed an automated, server-based toll free phone number that will be accessible to callers worldwide.