>Where is the evidence that there was even a single bot campaign - again, emphasis on bot - on the pro-net-neutrality side?

There's plenty:

>One 369-word comment supporting the Obama-era net-neutrality rules was posted on the FCC website more than 300,000 times. One of those was attributed to Gloria Burney, 87, a retired speech therapist in Los Angeles. She isn’t in favor of repealing those rules, she said, “but I never wrote that.”

You presented the "Chicago programmer" analysis as proof of bot activity. In the linked blog post, they say:

>A number of journalists have tried to trace where this message is coming from, and the best guess seems to be the Council for Individual Freedom. This was a mentioned in a recent Gizmodo article, where the organization claimed that they were running an ad campaign that directed users to a form, and they provided a screenshot of that form.

This is the same thing that many of the pro NN organizations were doing. Under what definition are you implying that a single message submitted 300k times wasn't a bot but the other ones discussed are?

Again:

>He said the FCC received more than 7.5 million comments consisting of the same short-form letter supporting the current rules from about 45,000 unique email addresses, “all generated by a single fake e-mail generator website.”

This is on average each email being used over 100 times. You think 7 million Americans:

A. Know how to use email generator websites B. Decided to all use a particular website C. Sent the same form letter while using said website and D. Happened to choose the same name as 100 other people on average?

This is just not a credible claim.

I'm sure you've got a perfectly innocent non-bot explanation for this as well:

>He said the FCC received more than 400,000 comments supporting the old rules “from the same address in Russia.”

> One 369-word comment supporting the Obama-era net-neutrality rules was posted on the FCC website more than 300,000 times. One of those was attributed to Gloria Burney, 87, a retired speech therapist in Los Angeles. She isn’t in favor of repealing those rules, she said, “but I never wrote that.”

This is consistent with my claim: "the pro-net-neutrality side organized virtual letter-writing campaigns, getting lots of actual human people to send in pre-written messages they agree with but did not author."

There is not enough detail in the WSJ article to distinguish between scenario 1, Ms. Burney having clicked on a thing that says "take action now to support net neutrality!" which sent a message under her name that she didn't write and no longer remembers, and scenario 2, Ms. Burney not having been involved at all, not wanting to have submitted a comment, but happening to agree with a message she never submitted. Both are plausible under the evidence, and we know there were sites that let people submit pre-written letters to the FCC and asked only for some contact info. I am claiming that scenario 1 is plausible and nothing in the article contradicts it.

What we know about Ms. Burney is that, first, she's familiar enough with the rules to have an opinion, and second, she "never wrote that." That's exactly what we'd expect from someone who sent in a pre-written comment.

If you want to argue that comments in scenario 1 count as "fake" comments for some reason, sure, we can discuss that, but first let's settle the claim you made, that these were "bot campaigns." Only in scenario 2 could it have been a "bot campaign."

> Under what definition are you implying that a single message submitted 300k times wasn't a bot but the other ones discussed are?

The definition that 300K real human people clicked a form and provided (or at least were directed to provide, and usually did provide) their own contact information, signing a pre-written message that they agreed with but did not author.

> You think 7 million Americans: A. Know how to use email generator websites B. Decided to all use a particular website C. Sent the same form letter while using said website and D. Happened to choose the same name as 100 other people on average?

Yes.

From the evidence we have, we know there were massive organized campaigns to get large numbers of Americans to submit, of their own agency, pre-written letters from the FCC. So it's not merely that it's theoretically possible, we have strong evidence for B and C, at least.

I'd have to dig up the report to see what happened, but a plausible scenario here for A is that one of these campaigns said something like "Your e-mail address will be public record, you may want to generate a throwaway address," and either it linked to this "fake e-mail generator website" or it was first on Google. In turn, for D, it's also entirely plausible that that website uses a low-entropy random email generator, such that it couldn't generate more than 45,000 email addresses and therefore had collisions.

What's your scenario? That a bot generated all these emails and used a third-party e-mail address generator website (instead of, say, generating its own emails)? Did the third-party e-mail address generator website have an API, or did the bot scrape it 45,000 times, decide that that was enough addresses, and then submit seven million comments?

> He said the FCC received more than 400,000 comments supporting the old rules “from the same address in Russia.”

E-mail address? IP address? Mailing address?

In particular, is your claim that a Russian bot campaign sent a bunch of letters to the FCC identifying itself as from Russia, knowing full well that the FCC takes into account only comments from America? How does that make any sense? I don't have an explanation, but I don't think you do, either, and you don't have enough evidence to call it a "bot campaign."

The standard of evidence you're requiring to call something a bot compaign is high enough that nothing counts.

So sure. It's theoretically possible that all the apparent hints of impropriety are just coincidences and lots of people just forgot what they did and so on. But then that goes both ways. It's not like the evidence is significantly stronger on one side or the other; if you're inclined to explain away everything, you can do so.

If you'd like, read my initial comment as saying something more along the lines of "there's some evidence of massive bot compaigns on both sides", or replace that with a weaker claim if you think none of this counts as evidence.

But motivated skepticism of the evidence on one side only doesn't help.

So: what evidence do you see that there was bot behavior on the anti-NN side that's so much stronger than any of the evidence on the other side? Everything I've seen can be easily explained away by the same standards you're applying.

The standard of evidence I'm applying is simply that evidence for X being true must not also be plainly compatible with X being false.

Here, for instance, is some evidence that meets that standard, from that very same article you love:

> He found a near-constant rate—1,000 every 10 minutes—punctuated by periods of zero comments, as if web robots were turning on and off. He determined many were from hacked accounts.

This pattern is either server-side (e.g., a server-side rate limit and an extremely flaky server) or client-side. If it's server-side, we'd expect to find it independent of the contents of the comment, we'd expect to have stories of people being unable to submit comments, etc. Unless we have those stories, then this pattern is plainly incompatible with actual humans submitting comments, even with the aid of some automation or pre-written letters. That is evidence that these comments were almost certainly not from actual humans.

No similarly strong evidence exists on the pro-net-neutrality side. "A bunch of comments claimed an address from Russia" is certainly suspicious and weird, I agree, but it is not plainly incompatible with humans (if it's IP address, bad geolocation; if it's e-mail address, Russian-run mail provider whose accounts are used by Americans; etc.). I am happy to grant with more detail about what's is meant there - which neither the FCC nor the WSJ has provided to us - there may be strong evidence, but it's not there yet with what you've posted.

If you want to revise your initial complaint to "Friendly reminder that some weak evidence exists in favor of small, isolated bot campaigns on the pro-network neutrality side and very large bot campaigns on the anti-network neutrality side, such that if all suspicious comments were disregarded in proportion to their likely untrustworthiness, the overwhelming weight of public sentiment as measured by this exercise would be firmly on the pro-network-neutrality side," then I have no remaining objection.

>Unless we have those stories, then this pattern is plainly incompatible with actual humans submitting comments, even with the aid of some automation or pre-written letters.

A site is set up to collect names for form letters. They don't submit it in real time; instead, they submit it in batches every few minutes.

This is at least as plausible as the various explanations you gave above.

In fact, if you read the linked blog post, they explain that this is how the API worked - you'd upload a CSV and everything would be uploaded in bulk.

And the sizes still don't add up. As mentioned, a single campaign for the pro-NN side had 7 million suspicious comments. That's bigger than any of the alleged bot campaigns on the other side. Overall, there's more suspicious comments for the pro-NN side. Whether the evidence is enough to definitively call them bot campaigns or not, that's still true.

And that's not even addressing the fact that WSJ contacted a portion of those leaving some of the form letters and found that 32% of one pro-NN batch said they never sent anything. The number is higher for the other anti-NN batches they checked, 60-80% range, but that's not the difference between massive and not-massive.

Assuming you trust the WSJ to have conducted the survey in an honest and non-leading manner, but I thought we weren't supposed to be trusting "biased media." The article starts off by portraying net neutrality as "pro-regulation" - I'd imagine a good number of people who are pro-net-neutrality would not have answered yes to, say, "Did you send a letter to the FCC favoring additional regulation on the internet?" despite having done so. We'd need to see exactly what the survey said and how they worded it.

They also haven't given you enough information to know the size of the pro-NN batch they investigated, so we still don't know that it's "massive."

>They also haven't given you enough information to know the size of the pro-NN batch they investigated, so we still don't know that it's "massive."

There's thousands of people who said they didn't leave the respective comments. That's enough to be statistically significant and to extrapolate across the whole sample.

There might be selection bias, as only a small percentage of people responded.

Sure. So how big is that population you can extrapolate over?

The WSJ article is carefully not saying, 32% of people who appear to have left pro-NN comments said they didn't make them. They said 32% of people in this batch did. They didn't say, as far as I can tell, how big the batch is. (Maybe I'm just misreading?)

There's a big difference between "32% of a thousand pro-NN comments are untrustworthy" and "60% of a million anti-NN comments are untrustworthy."

The WSJ, like any MSM outlet, is manipulating you through half-truths. Pay attention.

>There's a big difference between "32% of a thousand pro-NN comments are untrustworthy" and "60% of a million anti-NN comments are untrustworthy."

If you look on the sidebars, they reached thousands of people on both sides that disclaimed the comments left in their name. Given the abysmal response rate, the total size of the batches would have been hundreds of thousands at least for both sides.

>The WSJ, like any MSM outlet, is manipulating you through half-truths. Pay attention.

Look. I agree with you that we should be skeptical of such claims. But you seem to be applying that skepticism only to claims about one side.