Google: Security and Disinformation US 2016 Election | Hacker News Reader

Google: Security and Disinformation US 2016 Election

blog.google

25 pointsZhenya16 comments

If $4700 in ads could sway an entire US Presidential election, that has to be the most cost-effective use of money to achieve a foreign policy objective in the history of the world.

People believing it did is probably even more damaging than the ads themselves. Undermining confidence in our democracy == undermining our democracy.

It's far from that. Didn't Google report thousands of hours of YouTube videos created? Combine this with disinformation on Twitter, Facebook, and less transparent media like 4chan and Reddit, and the sheer amount of person hours spent impersonating Americans pushing one talking point or another can have a profound effect.

The problem with Correct the Record was, other than an extremely shortsighted objective, the fact that everyone knew about it. People who legitimately defended Hillary Clinton were assumed to be shills, to be "correcting the record" and this backfired tremendously.

Keep that in mind with these releases. The advertising dollars aren't the major news. It's the sock puppets and vote/conversation manipulation that are by far the largest component of the disinformation campaign. Those costs won't be public, either.

"We found 18 [Youtube] channels likely associated with this campaign that made videos publicly available, in English and with content that appeared to be political (These channels also posted non-political videos, e.g., personal travelogues). There were 1,108 such videos uploaded, representing 43 hours of content and totaling 309,000 U.S. views from June 2015 to November 2016. These videos generally had very low view counts; only around 3 percent had more than 5,000 views. These channels’ videos were not targeted to the U.S. or to any particular sector of the U.S. population."

That seems like a low count of channels, videos, and hours, does that include the content existing channels were paid to create? e.g.: the divisive messaging on Black Lives Matter, where the Internet Research Agency was found to have paid existing content creators to post messages for or against the organization.

> If $4700 in ads could sway an entire US Presidential election

That's a little disingenuous. IIRC, Facebook identified at least a $100k more on its platform, and they permitted finer-grained targeting than Google.

You're right that it's hard to believe stuff like this tipped the election outcome, but it's not impossible that it did. The sales pitch of targeted online ads is that they're cheaper and more effective than mass-market ads. Also, given the way US elections are conducted, the right pressure applied in the right areas could deliver an outsized effect on the outcome. Taken together the amount of money needed to sway a close election could possibility be less than one would expect.

> Also, given the way US elections are conducted, the right pressure applied in the right areas could deliver an outsized effect on the outcome.

True. But presumably the experienced veterans on Clinton's team also knew about it. It's not exactly a secret how US elections work, at least it shouldn't be the people trying to be elected.

$1B+ budget would have been enough to blanket all the right areas with ads day and night and even have money left over for people to walk door to door and ring doorbells.

> We found that two accounts linked to the Internet Research Agency spent a total of $4,700 on our platforms during the 2016 election cycle. This figure covers both search and display ads.

> Publisher products: We found less than $35 in AdSense and Ad Exchange revenue from ads on associated sites.

Then Facebook we found spent $100k on ads. Twitter apparently went directly to RT and implored them to target Americans with an ad campaign. Then made a brave and patriotic move to publicly tell them off recently.

Mueller was investigating for months and got Manafort for tax evasion and other corruption mostly around 2014 or so. That was when he was in bed with the Democratic lobby group headed by Tony Podesta, who apparently now is stepping down as well https://www.politico.com/story/2017/10/30/tony-podesta-stepp....

Clinton campaign: spent $1B+, had most of the large media conglomerates on her side, Wall Street, major tech companies sponsor and support her, debate questions passed to her ahead of time, best campaign managers, sitting president flying and campaigning on her behalf, and somehow some ads on Facebook and $5000 spent on Google overturned the result of the election?

Is this enough to convince everyone that there is probably no grand Russian collusion and they have been believing a conspiracy theory? Not really. After heading the wrong way for such a long time, it is hard to swallow ones pride and say "Oops, been driving the wrong way for about half a day, should definitely turn around". That's surprisingly hard to do, and is the reason people end up stranded in the desert or on an unmarked road in the mountains in a snow storm. Or another example would be the people in those doomsday cults, who surprisingly stay with the cult even after the end of the world date passes uneventfully. You'd think it would be pretty clear by then what's going on. But no, there is a new date set, and they stay and start believing ever more fervently than before.

The existence or not of collusion is independent on whether said collusion was ineffective or unnecessary.

The allure of populist retoric is very powerful, and this election showed it's a weak spot in the American system.

It's not a weakspot in the system. It's the system reminding politicians not to forget their masses.

Even Democrats are looking for ways to get back to their bread and butter blue collar voters[1] (the unionists who went for Trump and Sanders). The Dems (and neocon Reps) forgot about Joe and Jill sixpack and for decades advocated for business elites at the expense of these people.

These people have found a voice in populists like Trump and Sanders --they are not going away. It's the 80s with the rusting of American industry but now with a bit of revenge for all the exported jobs and job uncertainty and the advent of "sharing" economy and prominence of the "temp" workforce.

Of course, these neoliberal policies had some good, they are responsible for China's rise and bringing modernity and rising living standards as well as more affordable but also more disposable goods (not sure that's a good outcome in the long run --but certainly made some goods more affordable to the now struggling American middle class)

It's no secret neocons and Dems were enamored with free trade at the expense of blue collar voters who took stock and saw that their traditional Dem allies had sold them out.

[1]https://www.thenation.com/article/what-killed-the-democratic...

Weak spot? Its the administrative password.

> The existence or not of collusion is independent on whether said collusion was ineffective or unnecessary.

Am I misreading you or are you effectively saying that non-existent collusion can be as effective as existing collusion?

I've read it three times now.

(As usual I guess I should point out the sitting POTUS is and was not my preferred choice.)

Nah, I'm rebutting the argument of "the effect of any possible collusion was small, so no collusion existed".

I wonder if these companies can be trusted enough to provide accurate information, they might be tempted to downplay the amount spend on their platforms.

An example of this is Twitter. While they kicked RT off their platform, they did not come out openly saying - well we asked them to target Americans.

I feel that Google is trolling. $4,700 spent on ads? I know advertisers that spend that much per hour. This is a drop in an ocean. That’s 94,000 clicks at best ($0.05/click). But probably won’t get such low clicks. At 5% engagement rate you’ll engage 4,700 people.

The most interesting part of this is the timing - on the same day as the first indictments from Mueller, and a week before election day.

Reply on news.ycombinator.com