What’s new in tech
Twitter culls fake followers
Twitter began purging tens of millions of suspicious accounts from users’ follower lists last week, said Nicholas Confessore and Gabriel Dance in The New York Times, part of a major effort “to restore trust in the popular but embattled platform.” The cleanup takes aim at “a pervasive form of social media fraud.” Earlier this year, it was revealed that a Florida firm had sold fake followers to hundreds of thousands of users—including politicians, models, and actors—who wanted the appearance of social influence to bolster their careers, activism, or business endeavors. Many other companies offer similar services. Twitter’s cull had “an immediate impact,” with the platform’s total combined follower count dropping by about 6 percent.
That bot could be a human
Building a service powered by artificial intelligence is hard work, said Olivia Solon in The Guardian (U.K.). So hard, in fact, that some startups have decided it’s “easier to get humans to behave like robots than it is to get machines to behave like humans.” Many companies, including Facebook, have used people to help train AI systems to improve accuracy. But other firms “fake it until they make it, telling investors and users they have developed a scalable AI technology while secretly relying on human intelligence.” The business-expense management app Expensify, for example, has admitted it used humans to transcribe at least some of the receipts it claimed to process with “SmartScan technology.” Researchers say there’s an incentive for firms to hush up the human element of a service, because “people tend to disclose more when they think they are talking to a machine.”
FCC to overhaul comments system
The Federal Communications Commission wants to revamp its online public comment system “after millions of fake comments were posted about a recent FCC rule change,” said James Grimaldi in The Wall Street Journal. In a letter to Sens. Pat Toomey (R-Pa.) and Jeff Merkley (D-Ore.), FCC Chairman Ajit Pai said he’d asked Congress for permission to shift funds to pay for the overhaul. The senators had written to Pai in May to complain that their names were among those improperly used to post comments on Pai’s repeal of Obama-era net neutrality rules. Pai accepted the senators’ proposal that future commenters should fill out a Captcha, a tool designed to distinguish humans from bots.