Over a third of the nearly 22 million comments that poured into the Federal Communications Commission regarding its plan to repeal net neutrality protections included one of seven identical messages. More than half were associated with duplicate or temporary emails, including some 7,500 affiliated with the address "example@example.com." Dozens included references to the animated film Bee Movie, a film about a disillusioned worker bee that has become fodder for several popular memes. Roughly one million comments came from Pornhub.com email addresses. And more than 7,000 comments were submitted by a gentleman—or woman—named, simply, The Internet.
These bizarre facts and figures come from a newly released Pew Research analysis of every comment submitted to the FCC, which in turn illustrates just how chaotic the comment process has become in the age of automated electronic submissions. Since the FCC comment period opened in April, researchers have documented the apparent existence of bots that use natural language generation technology to masquerade as concerned citizens.
But Pew's report shows that even well-meaning individuals and advocacy groups attempting to flood the system with earnest comments have contributed to the crippling of the process.
"As public opinion researchers, we found it a little bit hard to really make sense of the public’s opinion on this issue," says Aaron Smith, one of the authors of the study.
Accord to Pew's review, only 6 percent of the comments were actually unique. The rest were submitted multiple times. Some of the most popular among them can be traced back to legitimate organizations that have mobilized their communities either for or against the repeal. In fact, by far the most popular comment—submitted an eye-popping 2.8 million times, accounting for a little more than one in 10 entries—was a pro-net neutrality missive, promoted by television host John Oliver and featured on the website battleforthenet.com. Three of the top 10 most popular comments, meanwhile, were anti-net neutrality messages promoted by the Taxpayers Protection Alliance.
While both sides of the fight used this form letter technique, the anti-net neutrality side may have had more success. Of the top seven most-repeated messages, six included anti-net neutrality sentiment.
The organizations using these tools to mobilize mass audiences may be driven by good intentions, but the research suggests that neither they nor the FCC had the verification mechanisms in place to ensure that the system wouldn't be abused. The FCC's dataset includes a field indicating whether email addresses had been verified, Smith explains, but for the vast majority of submissions, that field is blank. In fact, only 3 percent are definitively marked as verified. As a result, more than 9,000 emails provided to the FCC didn't even include the @ symbol, which is, of course, required to create an email address. Another 8 million submissions included phony, temporary email addresses purchased online from sites like FakeMailGenerator.com. The researchers detected them by matching the email addresses attached to comments to the set list of domain names FakeMailGenerator.com provides.
It's unclear whether this was an effort perpetrated by one side or the other, or whether the commenters themselves were fake. "There are any number of legitimate reasons why someone might want to use these types of accounts," Smith cautions. "But in a holistic sense, the use of these temporary emails, taken together with the large share of comments that utilized duplicate email addresses, nonfunctioning email addresses, or who simply left the email field blank, does make it challenging to determine who or what these comments are coming from."
The Pew researchers detected other unusual behavior, like the fact that on June 19, nearly 500,000 comments were submitted in a single second. In that case, nearly all were identical and associated with battleforthenet.com, suggesting its organizers bulk-uploaded all of the comments to the FCC's site at that time. But on May 24, they found more than 86,000 comments submitted in a single second. They all conveyed the same sentiment, but this time, the language different slightly, following a pattern that other researchers recently told WIRED may have been generated by bots.
Smith and his team stop short of suggesting what the FCC might do about this issue. There's nothing wrong, after all, with advocacy groups giving their supporters a little assistance in making their voices heard in Washington. That's their job. And under the current leadership, it's unclear whether the FCC, so driven to overturn net neutrality protections, has the motivation to do anything about it at all.
Smith says the goal of the research is simply to "show some of the challenges that researchers face as they try to analyze this type of data and highlight the way in which these wide-scale campaigns to influence comment processes are much easier to do at a scale than a few years ago."
Now that that ability is being exploited, the earnest groups and individuals who want to have a say in their government's rules may need to try a new approach to cut through the noise.
