The next day, after work, he decided to visit a friend. As soon as he had sent the first text message, the system had predicted several possible routes that he would take. As he walked the twenty blocks to Meilin’s apartment, cameras scanned the streets, facial recognition algorithms detected his presence, and a map of his route was updated in real time. As he passed from the range of one camera to the next, the system tracked his exact path, logging the times and geolocations. Over the next few days, he exchanged messages and visits with a few friends (acquaintances, really) and coworkers. Because of their link with Chu, however tenuous, each of these persons also became the subject of automated assessment. None of this had required any human intervention - it had all been done automatically. An investigative agent, notified of the new activity, had performed some minimal supervision, and made minor changes to the surveillance tasking here and there, but these were mostly busy work. The agent’s caseload was large, and he didn’t have much time to spend on any one case.
One of Chu’s friends heard his frustrations out, and responded with an animated speech about the oppression of laborers, the perversions of ‘communism’ under the current system, and some other things that Chu didn’t really understand. This friend mentioned meetings of a group he was helping to organize, encouraged Chu to attend, and instructed him in how to use a special app to discuss them. Talk of secret meetings made him very nervous, but his friend was so passionate, and seemed so understanding of his plight, that he didn’t have the heart to turn him down. Living in the city had made him so lonely, he was glad to be asked to participate in something. He decided to be noncommittal, and to put off going to any meetings so he wouldn’t be actively involved. They exchanged a few messages, but the encryption was purely a window dressing. Superficially, the app offered encryption, but all companies were required to provide the government a backdoor to decrypt messages. The messages were processed using sentiment analysis and other newly developed methods which had not been published in the open literature, and their content was flagged as likely being antisocial and possibly subversive.