So Who Posted All Those Fake FCC Comments About Net Neutrality?

After nineteen months, a couple Freedom of Information Act requests, and a (still-ongoing) lawsuit against two federal agencies, I’m pleased to announce the start of an answer to a question that has lingered for far too long:

Who bombarded the Federal Communications Commission with all those fake comments about Net Neutrality back in 2017?

Thanks to the efforts of my attorneys Josh Burday and Matt Topic of Loevy & Loevy, the FCC has been forced to release 784 .csv (comma-separated value) files submitted to their Electronic Comment Filing System (ECFS) between April 28 and May 28, 2017.

These files — which can be opened, edited, and saved in spreadsheet programs like Microsoft Excel — collectively contain nearly 3.4 million rows filled with name, address, and comment data corresponding to the submission fields in the FCC’s ECFS.

Given the nearly 22 million comments accepted by the FCC in total, this data alone constitutes around 15% of the total comment submissions…not all of which were accepted (more on that later).

The FCC also provided a master log .csv file, containing the submission time of each file, the file name, and submitter’s email address:

You can view and download this data here, but I will be including screenshots of the relevant points throughout.

With that in mind, here’s what I found:

Single “bulk” comments

The vast majority of these .csv files:

1. contained only one comment

2. contained a pro-Net Neutrality comment

3. were submitted with a unique email address, and

4. were submitted around May 8–11, 2017 when the overload of web traffic attributed to comedian John Oliver’s gofccyourself.com campaign (falsely claimed by the FCC as a “DDoS attack”) temporarily overloaded the FCC’s comment system. In fact, of the 550 or so “valid” single comment .csv files, 496 of them were uploaded on May 8, 2017 alone.

Many of these single comment .csv files, however, were sloppily reformatted from the FCC’s original .csv template.

For example, plenty of these single comments contain the name “John Doe” and the other filler info from the original FCC .csv template. Many of these comments also contain the language “Dear FCC, I am writing you today because I am hoping that my voice actually counts.”

While this pattern is clear across the majority of the .csv files, the comment volume they contain is practically nothing compared to the 5–6 clearly distinguishable pro- and anti-NN email accounts that bulk submitted most of the 3.4 million comments through May 28, 2017.

Fight For The Future

While there are three instances of a personal email address uploading bulk pro-NN comments in the hundreds or thousands, three clearly identifiable groups — Fight For The Future (362,448), Free Press (42,408), and Demand Progress (58,708) — posted a combined total of less than half a million comments through May 28, 2017. Interestingly, all three groups only did so on May 11, 2017.

To an extent, all these groups also uploaded duplicate comments.

In the clearest example of this, the logs show that Fight For The Future uploaded every .csv file name at least two or three times across three distinct timeframes during May 11, 2017. Though there appear to be slight differences in comment totals across the duplicate file names, FFTF only uploaded 16 unique .csv files, but a total of 40.

The FFTF .csv uploads also had consistent formatting problems that can be seen in the final FCC ECFS record. Take, for example, this submitted comment from “Taylor Blawusch” in which Column M — which is where the actual comment is supposed to go — contains the date entered for the “Date Received” Column L instead:

The final result?

Similarly, there are instances where FFTF divided the comments between multiple columns in the spreadsheet, but only the text that made it into Column M got posted. For example:

Many of the names in the FFTF comments also have multiple comments about “Restoring Internet Freedom” under their name, including comments submitted after May 11 that wouldn’t have been included in the duplicate files:

Given the widespread popularity of Net Neutrality, and FFTF’s forefront role in leading in the pro-NN/anti-“Restoring Internet Freedom” public commenting campaign, its plausible that most to all of these comments are from real people.

However, the irony is that FFTF’s formatting errors resulted in the majority of these likely real people’s comments not being submitted in full. Even with the boilerplate language they provided, the end of their sentence “not the narrow interests of Internet Service Providers” is consistently cut off.

Demand Progress

Demand Progress submitted one .csv file and every comment said: “A free and open internet is critical for Americans to connect with their friends and family, exercise their freedom of speech, and create innovative new businesses. In 2015, the FCC established strong net neutrality rules to protect the free and open internet Americans depend on. Please reject any plan from Trump or his FCC Chair to roll back net neutrality rules and open the door to a corporate controlled internet.”

There are a few instances where duplicate names with the same address and slightly different listed email addresses show up, but otherwise the DP submissions appear to have unique name/address/email combinations.

Free Press

Of any group, pro- or anti-NN, the individual comments within Free Press’ four .csv file submissions looked the most like a plausibly real collection of varied individual responses to a given issue:

According to Free Press’ Tim Karr, the group used online organizing tool ActionKit, which allowed their user base to fill out fields as they pleased (hence the “xxx” frequently appearing in the address column), match entry fields “to some but not all of the FCC’s categories,” and export data as a .CSV file. Hence, why the Free Press comments do not contain individual email addresses, unlike nearly every other comment.

However, when I combined all four files and sorted them alphabetically by name, there were significant numbers of fairly distinct duplicate names, but with different addresses and posting different (or the same) comments.

But in one example illustrating the complications with determining what is “fake” or “duplicated”, two comments were uploaded by Free Press under the name “Steve Wozniak”.

Accordingly to Karr, these are, in fact, two members of Free Press who aren’t the Apple co-founder (who himself is pro-NN). Generally, he said that because Free Press sends many emails to the same or similar lists of people, “it’s not unusual over a stretch of time for one person to comment more than once on an issue.”

In a curious twist, Karr also claimed by email that part of the terms of his organization’s subpoena by the New York Attorney General’s office is that all communications with outside parties about their comment submission to the FCC needed to be forwarded to the NYAG — including the very email I was reading, and my own prior correspondence with Karr.

While the NYAG’s office did not respond to my request for confirmation, it is reasonable to assume the 13 other groups allegedly under subpoena are currently under the same terms.

MediaBridge

The majority of the actual comments coming from .csv uploads through May 28, 2017 (but falsely attributed to bots) were pro-“Restoring Internet Freedom” and anti-Net Neutrality.

Of the anti-NN bulk submitters, only MediaBridge is clearly identifiable as an originator.

As the largest single uploader of bulk .csv comments (1,856,552) during this period, there’s a curious divide between the comments submitted by the email address of MediaBridge founder (and former Project Veritas Executive Director) Shane Cory.

Some of his uploads consist of thousands of names and a single comment, like shown:

The majority of the MediaBridge submissions, however, fit the mold of what data scientist Jeff Kao (now with ProPublica) described as generated “mad-libs” comments:

Sure enough, Cory told me by email that these comments came from an “in-banner form collection with a rotating letter” with “6 septillion possible versions” to avoid the appearance of a form letter. While he would not name his client, he claimed that they also wanted “a standard, non-dynamic letter later in the campaign” in order to “make sure his organization was on the public radar with a form campaign.”

However, Cory did not provide visual evidence of a rotating letter or banner ad campaign. Furthermore, instead of one “non-dynamic letter”, MediaBridge’s “not mad-lib” comment submissions consist of five different messages posted in bulk. This includes language identified by Pew Research Center as entering into the FCC ECFS 90,458 times at the same time.

Or in other words, “batch-d-5.csv”:

This is a good time to point out that the Pew study reasonably — but falsely — attributed all these simultaneous submissions to bots, much like many media outlets.

In fact, they are likely the result of each .csv file being manually uploaded to the FCC one-by-one into the Box.com-powered widget by a real person…regardless of whether the individual comments they contain were “real”.

Like the pro-NN comments in general, there are duplicate name submissions by MediaBridge:

Like FFTF, there are entire duplicate files that were re-uploaded at a later date. To this, Cory claimed that because he wasn’t sure which of his initial comment batches were uploaded and accepted, he resubmitted many of them just in case.

However, unlike FFTF’s duplicate files, all of the MediaBridge .csv submissions are under different names with two different conventions. For example, the 36154 comment batch uploaded on May 18, 2017 under the name “file-h” was uploaded again on May 26, 2017 under the name “Batch-A7”:

It’s possible the Box.com submission widget screwed up during the mid-May 2017 chaos surrounding the FCC’s ECFS in general, and it’s true there was a “sunshine period” between May 12–18 where the FCC was not accepting new comments in order to catch up with the previous submissions. But if the original files were being re-uploaded as-is “just in case” as Cory claims, we would potentially be seeing duplicate file names in the submission log, like with FFTF.

Instead, it appears every MediaBridge .csv file submitted on May 18, 2017 with a name starting with “file-“ was resubmitted within the week under a new name starting with “Batch”.

Also, between May 18–28, Cory was the only bulk .csv submitter, other than…

The Mystery Anti-NN Uploader(s)?

The last “two” major bulk submitters are anti-NN, but are not readily identifiable — and may just be the same person.

Between May 9–10, 2017, ten .csv files were uploaded by fccfreedom@hmamail.com.

These consisted of one of two messages, starting with either:

“In 2015, President Obama’s FCC passed rules treating the Internet as a government regulated public utility for the first time in history. Those pushing hardest for the new rules were Silicon Valley monopolies like Google and leftist globalists like George Soros…”

or

“Obama’s Federal Communications Commission (FCC) forced regulations on the internet that put the government, and unaccountable bureaucrats, in control…”

Starting two hours after the last fccfreedom upload, the exact same file naming conventions and bulk comment language begin to appear under the email address esmisc@mac.com:

However, starting May 18, esmisc@mac.com stopped uploading bulk comments with the first message, and instead uploaded comments with a similar, but new message starting with:

“In 2015, wealthy leftist billionaires and powerful Silicon Valley monopolies took the internet out of the hands of the people and placed it firmly under the thumb of the federal the government, monopolies like Google and global billionaires like George Soros.”

As you can see, the file names either start with an F/”FOI” or T/”TPA” followed by a date, and often, the total number of comments in that file. With a few exceptions, the “T” or “TPA” files contain all comments starting with “Obama’s Federal Communications Commission (FCC) forced regulations”, whereas the “F” or “FOI” files mostly correspond to the old and new versions of the Silicon Valley/Soros conspiracy theory language, respectively.

What “FOI” stands for is anyone’s guess, but “TPA” most likely stands for Taxpayers Protection Alliance, a Koch-funded group whose talking points were found and recognized by the Pew Research Center study in many of the FCC comment submissions, and who are also allegedly under investigation by the NYAG’s office.

It’s unclear if the two email addresses are separate operatives for the same client or the same person. If it’s the latter, it’s especially unclear why someone would immediately switch from a burner email address to using their Apple ID. Either way, the combined email addresses posted around 1,023,398 comments through May 28, 2017.

Like with the others, the .csv files uploaded by these email addresses contained duplicate names, addresses, and emails, but often had differing email addresses that didn’t remotely resemble the alleged real poster’s name:

Most notably, this mac.com address posted one of the most famous fake comment examples of all:

Given that the late actress Patty Duke died before the “Restoring Internet Freedom” commenting period, these comments are the most likely batch to have been completely fabricated.

But who would be dumb enough to mass upload fake comments to a government website under their own personal email address?

Let’s start by assuming the NYAG’s alleged subpoena list includes every major group directly (or indirectly) involved with the bulk comment submissions. We can obviously rule out Free Press, Demand Progress, FFTF, and MediaBridge for these particular batches.

That leaves CQ Roll Call (who I first reported asked the FCC for advice on how to make millions of comments look real), Christian Coalition founder Ralph Reed’s Century Strategies, ad agency LCX Digital, Broadband for America, Center for Individual Freedom, and the obvious candidate, the Taxpayers Protection Alliance themselves.

There’s also Vertical Strategies, a marketing/political consultancy firm. While this could be a stretch, they previously employed Ethan Eilon, a “Political professional and self professed apple nerd” who was also involved with an anti-Net Neutrality porn parody video from 2015.

Whoever it is, evidence that their own email address was used to post an extremely high-profile example of identity fraud is now a part of the public record.

Not accepted/missing comments

For all the FCC’s talk about needing to accept (if not necessarily consider) every comment that came in, there’s one thing they declined to mention — they didn’t accept all of them.

There, of course, were instances of poorly formatted .csv files I found that wouldn’t have made it in no matter what.

However, I found six files with the name “FCC_NetNeutrality Advisement 17–108”, but uploaded using six different — and sometimes interesting — email addresses:

All of them consisted of the comment “Dear FCC, If my voice does indeed count then I want to stand up for Net Neutrality and support freedom over the internet instead of being controlled by Corporate Greed” 601 times with individual names:

When searching for any of these 601 names listed, none of them appear, or if they do, not with this comment:

So either the FCC made the same mistake 6 times, Box made the same mistake 6 times, or the FCC actually did choose to regulate and deny .csv file submissions — undermining their entire position on not regulating the comments.

Conclusion

Whether you want to blame the FCC, Box, or the bulk submitters, its clear that this public commenting process was not only a failure, but a failure on multiple levels by multiple parties for multiple reasons. At the very least, this .csv bulk commenting option turned out to be far from foolproof, adding even more of a mess to an already fraught process.

Admittedly, we still don’t know 100% for sure “who faked the comments”, and I’ve very likely raised more questions than answers.

But once again, this is only a fraction of the data. My FOIA lawsuit (which is currently on stay due to the government shutdown) is still ongoing to obtain logs related to the FCC’s API comment submissions. Plus there are likely 5 ½ months worth of additional submitted .csv files that can now be Freedom of Information Act requested from the FCC, thanks to yours truly.

But if this chunk of data is any indication, the FCC still has a lot to answer for.