How Google Could Rig the 2016 Election | Hacker News Reader

How Google Could Rig the 2016 Election

politico.com

33 pointsgrej29 comments

I'm really not concerned with Google's ability to "rig an election", when the mainstream media's editorial choices already exist and get far more mindshare. He gives the red herring of there being multiple, eg, television stations with lower market shares than Google, but all of those theoretically competing media sources (yes, even the dreaded Fox News) have a very narrowly defined set of allowable opinions and they push them relentlessly.

And further, if minor changes to news composition are sufficient to "rig an election", maybe we should start talking about whether democracy is actually a good decision mechanism.

One mechanism that I find very striking, is when journalists make the choice between stories like "X says Y" vs "X refuses to apologize for saying Y".

Pff, voters. Wake me up when Google is manipulating the electoral college, that's where the action is. Manipulating voters is the explicit raisin d'être of every news organization, politician, and super pac in existence. One more player in that field isn't changing anything.

I feel like this article is going from a well-supported and accepted point—that search engine results can influence our decisions—to then just belabouring the point and making dark allusions to this "powerful" effect and the "possibility" of Google rigging an election.

Don't forget the nefarious possibility of a rogue AI governing us all from the shadows! Or how those pesky Googlers deleted the election data from the Internet but NOT before our hero and his colleagues got it safe! Surely we must sign another huge grant for this man to save the planet.

First off, the WiFi thing. Ffs. The difference there is literally typing "-s 0" or omitting that parameter vs "-s 64". Hardly anything evil or firable or criminal. Making them seem bad because they didn't fire someone for a trivial mistake is silly (and anyways it was public broadcasts, not that it totally excuses it but come on).

Second, Google's entire point of existence is to give an opinion on search results. Full stop. If Google decides that people searching for Clinton want to see positive articles, that's their call. Again, that's their entire product. People are trusting Google's opinion on search results.

And calling Google's response meaningless? Google's saying they're screwed if they deliberately betrayed users. And what they aren't saying is that if they ever were caught really manipulating results, they would be hit with so much regulation that Bing would have a chance.

>Second, Google's entire point of existence is to give an opinion on search results.

Nope, their point of existence is to sell ads. Showing good results is how they create value, and you are just stretching the meaning of "opinion".

>they would be hit with so much regulation that Bing would have a chance.

This is just naïve. Google is one of the biggest companies in the world and consequently their lobbying power is huge. Besides, I don't even know how they could regulate it. Their algorithm will always be a trade secret and there is no regulation that could ever force them to publish it.

You're right about the ads. I should say the entire functionality of the search product itself. As far as "opinion", while you might call it a stretch, it's still quite valid and true. There isn't any standard definition of "relevant".

As far as regulation, eh, I'd love (in a perverse way) to see what the EU comes up with. They could assign officers that get secret reports on algorithm changes. They could mandate that Google justify, in secret, every change, with scrutiny on anything that doesn't serve a clear purpose. They could make it a criminal offense to intentionally change results to skew elections. A government isn't just gonna shrug and go "oh well, trade secrets".

There's no way that Google wants more people thinking "monopoly" and "fair search results".

Yeah, it will be interesting to see what comes out of Europe. Personally I think that regulators don't stand a chance. For example, they passed a Google (News) tax in Spain, and Google closed the Spanish Google News page. Now Spanish newspapers are begging Google to bring back Google News.

While there are many situations where regulation is important, I'm not so sure about governments secretly interfering with how media report on electoral candidates. The agency responsible for this would have to be absolutely guaranteed completely independent from political branches of government and established parties.

We don't accept this kind of interference with newspapers and TV news. Why would it be acceptable for search engines?

I find it strange that the article claims that Western Union's manipulation of coverage influenced Hayes' election, since as any worthwhile student of American history ought to know, the election of 1876 was marred by far, far worse facts. Hayes actually lost the popular vote, but this needs to be tempered by the fact that both sides engaged in voter fraud and intimidation (e.g., Democratic ballots being printed to fool illiterate voters (i.e., Southern blacks) into thinking they were voting Republican). Partisan electoral commissions duly responded by throwing out ballots to make sure the electoral votes went the right way.

The Democratic House balked at accepting these results, while the Republican Senate claimed it didn't matter what the House thought when it came time to officially count the votes. This crisis was resolved by creating a commission, but the Democrats decided to try to buy the vote of the only true independent, who responded by showing surprising integrity and immediately resigning, which left the commission with having to choose the most supposedly impartial Republican member of the court to cast the swing vote, which was finished only two days before the inauguration.

In the face of all of that, claiming that Western Union support was decisive in landing Hayes the presidential office is intellectually dishonest.

> So they fired him, right? Nope. He’s still there, and on LinkedIn he currently identifies his profession as “hacker.”

Oh no, a "hacker". He's definitely nefarious then.

The Google Headlines on my new Android phone is basically a mechanism for seeding emotional contagion and topic sharing at work, home, etc. This should not be underestimated. What happens when you push a particular site's article about a particular topic, at a particular time of day, to millions of phones? Who seriously is not worried about the susceptibility of this mechanism to manipulation?

I'm worried, but only slightly more worried than I already am about the effects of mass media on societies.

Articles like these often give me mixed feelings. On the one hand, it's good to point out these specific dangers. On the other hand, it strikes me as counter-productive to focus or spend energy on this 'hypothetical' situation when the current media landscape is already way too big and bad an influence on our democracy.

It's a bit like discussing Obama's track record ad nauseam when it's quite likely that he himself has limited influence, and the problems are deeper and broader. It's a case of focusing on a symptom at the expense of fighting the root cause.

(that said, I'm also conflicted because it could also be said that it's better to get going, fight a symptom, and see where the path leads, rather than sussing out the root cause and then feeling powerless to do anything about it)

Google's response: http://www.politico.com/magazine/story/2015/08/google-2016-e...

Google is hardly the only one with this kind of power. News media have been doing this for ages, and are doing it far more explicitly. How many news stations in the US are still neutral and unbiased?

Worth noting, for example, is the relative lack of coverage for Bernie Sanders in mainstream media, despite his name being all over the internet. Could that be related to Sanders wanting to break up big media conglomerates?

Google seems a lot less politically motivated, and seems to be under a lot more scrutiny, than most news media.

Ok, the general idea of the article totally makes sense to me. But the argument doesn't work out for me:

> That high volume of search activity could easily have been generated by higher search rankings for Modi.

Is it just me or is this article confusing the connection between "search ranking" and "search activity"? What influence does the order of the results for a candidate's name have on how ofter people search for this candidate's name?

Can anyone explain what I am missing?

As far as I can tell the argument goes like this: people are only likely to click on the number one or number two search result for a given search. Pushing favorable news stories about a candidate to the number one or number two spots (instead of neutral or negative stories) influences public opinion positively in favor of the candidate.

The response by Google is not exactly convincing.

What would you have them say? It's the obvious answer: if we start rigging elections and get found out, we lose our customer's trust and hence their business. We have a good thing going here and value that too much to screw it up by "being evil".

I think losing their customer's trust is the least of the problems they would experience.

It's no answer at all, and it's certainly not a foregone conclusion. They could easily just blame it on a rogue employee etc, and there is absolutely no reason to believe that they would lose business, since competitors would be equally suspect.

If they wanted to actually rebut this they could easily say something much stronger such as: "This would be absolutely against company policy, and we have process in place to safeguard against it which we publicly document."

The problem is that the search results almost certainly do reflect the company's biases even without any intentional rigging, so they can't answer that way.

The answer reads like this:

  We absolutely do not have the capabilities to do this and anyone who says otherwise is a liar!

What follows is a list of reasons for why they don't have an incentive to do this rather than reasons that they can't do this.

At one point Google was eliding several prominent conservative personalities, amongst them Pat Buchanan, from its autocomplete.

I'm assuming the downvotes are because claims like this are very common, and without some kind of presentation of evidence, and discussion of the explanation (if any) given by Google, they just become background noise in any discussion.

As any fan of the Foundation series knows, the true power of mass data is not in chasing down individuals, but in being able to statistically predict and manipulate the group average.

Google are not the only player in this area though, I suspect that Facebook can also reliably influence enough people.

> Google are not the only player in this area though, I suspect that Facebook can also reliably influence enough people.

Facebook have published research in this area of emotional manipulation through their news feed: http://www.pnas.org/content/111/24/8788.full

> We show, via a massive (N = 689,003) experiment on Facebook, that emotional states can be transferred to others via emotional contagion, leading people to experience the same emotions without their awareness. We provide experimental evidence that emotional contagion occurs without direct interaction between people (exposure to a friend expressing an emotion is sufficient), and in the complete absence of nonverbal cues.

Old school FUD! I haven't read anything like this in awhile.

they greatly underestimate the complexity of preference algorithms. Facebook is said to have more than 100,000 variables in their news feed algo; I can't imagine what's under the hood at Facebook. In a system that complex, making gross decisions trying to compromise one particular set of results without compromising everything else is not just idiotic, it's impossible.

Also, I enjoyed the yellow journalism gems in there, such as phrasing like "Google refuses to apologize for" etc.

BUT just because the idea is stupid and unworkable, does not mean we should discount the article. It will resonate with a lot of non-engineers, and it will be repeated. So I would advocate for some kind of "common carrier" laws for enormous information brokers like Google, Facebook, and Twitter. Never mind that actually manipulating the results would be prohibitively difficult and dangerous to the health of the company, and proving manipulation would be effectively impossible. If such a legal framework makes people who don't understand complex engineering sleep better, I support it.

>not just idiotic, it's impossible.

I am not even that great of an engineer but I could probably sit down and in a couple hours come up with a hundred different ways to do this.

Reply on news.ycombinator.com