infrastructure

REGULATION: The Ethics of Sovereign Technology

f96f8b7a-83dc-4582-866f-67aa3f6afae9[1].jpg

The Chinese tech companies sit comfortably between media, software and finance. No distinction needs to be made between using someone's social media data, search history, shopping habits, education and financial track record -- all of these data points flow into massive AI power-houses with half a billion users, inside Alibaba, Baidu, Tencent and others. Now, imagine if your Facebook friends and Google searches and Amazon shopping and Visa purchases determined if you could get a student loan to go to university. China's social credit system will do just that, reports Futurism.com while referencing an infamous Black Mirror episode that explores a dystopian view of this concept.  While countries like the US certainly struggle with systemic bias in the commercial activities of free participants, at the least such bias is not put into software by the federal government and used to determine access to services. 

Or is it? Consider that a regulatory agency, the Federal Communications Commission (FCC), is working to remove net neutrality rules in order to allow Telecom companies to meter how and where Internet traffic flows. While Comcast may not turn off access to some particular site, they could in theory tier the speed of the Internet according to economics, politics or whim -- such that, for example, Goldman Sachs or Amazon load quickly, while Crypto Kitties or Telegram load slowly. And loading time is a major determinant of consumer behavior and information exposure. We can't trust the Telecoms to not extract economic rents. Therefore, this policy choice will strengthen the ability of entrenched commercial interests to determine what people have access to, and consume.

And what if the rationale for such policy decisions is not even driven by the sovereign, or the collective will of that state's people? Data Scientist Jeff Kao deed a deep dive on the FCC comments using natural language processing analysis. In the best case, at least 1.3 million of the pro-repeal comments sent to the FCC were automated. In the worst-case, only 800,000 of the 22 million submitted comments came from real people, with 99% of those opposing the repeal. In the graph below, the height of the bar shows how often a campaign was repeated on an exponential scale. The color Red is associated with repealing and the color Green with keeping net neutrality. The clustered Red middle suggests the work of spambots that can generate slightly different language with the same meaning. The long Green tail is likely to be written by real people, though the first top bar shows a form letter repeated 7.5 million times. This cuts in every direction.

cc11a4a3-20dc-4f83-b1a5-ff8decf1b58f[1].png



And of course, other examples of propaganda bots are well documented -- already in use by 30 governments, with election impact in 18 countries. According to a recent report from Freedom House, Internet freedom is declining on a global basis. The mental stretch from a government controlled social credit system to global information warfare over national policies is not as unlikely as it may first appear.
 

Source: Freedom HouseJeff Kao