On November 23, Jeff Kao, a data science student from San Fransisco, California, ran natural language algorithms on the Federal Communications Commission (FCC) comment thread regarding the repeal of Net Neutrality. The 2015 law was signed in by former United States President Barack Obama and imposed increased government oversight of broadband traffic.
“It meant that Internet service providers became treated as public utilities and were forbidden from blocking or slowing rivals’ content. The rules also applied open-internet protections to wireless services for tablets and smartphones,” writes Gerry Smith for the Washington Post.
However in May, current FCC chairman Ajit Pai, who voted against the 2015 law, took the first formal step toward dismantling the net neutrality rules. The Federal Communications Commission has scheduled a vote for Thursday on whether to abolish the rules.
Kao created the algorithms due to the comments by New York Attorney General Schneiderman.
“Schneiderman estimated that hundreds of thousands of Americans’ identities were stolen and used in spam campaigns that support repealing net neutrality,” Kao wrote on his blog.
His algorithms recognized over 1.3 million possible fake comments based on patterns found within each comment stream.
Jeff Kao’s key findings presented on his thread:
- One pro-repeal spam campaign used mail-merge to disguise 1.3 million comments as unique grassroots submissions.
- There were likely multiple other campaigns aimed at injecting what may total several million pro-repeal comments into the system.
- It’s highly likely that more than 99 percent of the truly unique comments were in favor of keeping net neutrality.
This presents the possibility that supporters of United States net neutrality regulations, initially put in place as the Open Internet Order in 2015 as part of the Title II rules established in the Communications Act of 1934, may be drowned out by machine-based comment generation systems that skew the true majority of comment opinions in favor of a minority.
An investigation by the Wall Street Journal echoed Kao’s findings.
The Journal found instances of fakes that favored antiregulation stances but also comments mirroring consumer-groups’ pro-regulation talking points, posted without permission of people whose names were on them.
However, the FCC says it will go ahead with its vote this week nonetheless. (NPR)