Project Veritas Exposes Google

[MisterSwig]

Project Veritas released a new video in which a Google employee exposes the company’s plan to suppress right-wing voices and affect the 2020 election against Trump. Also, he talks about “algorithmic fairness,” which is Google’s program to condition users into a left-wing worldview. For example, they manipulate search auto-completes in favor of their progressive agenda.

 

Google should have the right to control content on their websites. But there is a serious question about whether they are defrauding users who don’t know about the company’s biased manipulations of information.

https://www.youtube.com/watch?v=re9Xp6cdkro

 

[Eiuol]

I mean, I don’t like the huge emphasis on that algorithmic fairness thing. I don’t think it’s a good thing. But this video I think really exaggerates the issue into some dark and sinister manipulation. Plus the auto complete search terms is a different process anyway, which aren’t any more than suggestions anyway.

An effective search require some sort of valuation and ranking, which on some level requires a company to use its own values. You shouldn’t fault a company for having values. Sure, judge its values, but you seem to be saying that Google is wrong to express any values. It’s not like they are deleting search results. If they were deleting results, I think you would have a point.

1 hour ago, MisterSwig said:

Google’s program to condition users into a left-wing worldview

This is not how conditioning works. That’s just a scare word. Being conditioned requires a reward (or disincentive). All you have here is Google offering search results with nothing to lure you into specific links.

[Grames]

5 hours ago, Eiuol said:

All you have here is Google offering search results with nothing to lure you into specific links.

If that had no effect why would Google bother to do it?  I hypothesize that just maybe Google knows how their autocomplete feature is used and relied upon more than you or I do.

Why can’t objectivity be a value at Google?  I mean, anything other than the most frequently searched terms being auto-completely is a deliberate sabotaging of their own service and clearly non-objective in the ordinary arithmetic sense of objectivity.

[Eiuol]

4 hours ago, Grames said:

If that had no effect why would Google bother to do it?

I’m sure it makes people more inclined to look at those search results simply because it’s there. What I mean to say is that I don’t see them providing an incentive, nothing to increase or manipulate the appeal. If they did, that would be conditioning. For example, loot box are something common in videogames these days. You keep buying them until you get the item you want. A lot of people don’t like them. But they work because of conditioning. There’s a behavior and reward for that behavior, associated with that game specifically. This is what I think of when I think conditioning.

5 hours ago, Grames said:

Why can’t objectivity be a value at Google? 

I think it should be, and I fault them for not holding that as a high value. Although this doesn’t have to be sinister, or some manipulative agenda, it’s still bad. I don’t use Google nearly as much as I use to these days, partly for privacy reasons, but also because I don’t get the sense they are as objective as they used to be. I guess the video gave me more reason to think that my sense of things was really what was going on after all (less concern for objectivity).

[MisterSwig]

YouTube removed the video, so you’ll have to find it at the Project Veritas site if you’re interested.

14 hours ago, Eiuol said:

Plus the auto complete search terms is a different process anyway, which aren’t any more than suggestions anyway.

Huh? It’s part of the “algorithmic fairness” program. It’s the first example given in the video. I ran this search (and others) myself to verify it. As stated by the whistleblower, instead of being based on actual user searches, it’s pulling auto-complete results from the “source of truth database.” Frankly, this was a little disturbing to me. I have suspected that something was weird about these auto-completes before, but it never occurred to me that they weren’t based on user searches. I don’t use Google to have their political agenda imposed on me. I want and expect to see popular search results. If not that, then let me easily filter the auto-complete results the way I want.

14 hours ago, Eiuol said:

Sure, judge its values, but you seem to be saying that Google is wrong to express any values.

I’m saying their “values” are actually disvalues. They’re being dishonest with the users who have come to expect a certain level of neutrality. I understand that bad users try to manipulate the search results by flooding the Internet with their garbage. The answer, however, isn’t for Google to then manipulate the results with its own garbage philosophy. But if it does, it should make clear that it’s doing this.

14 hours ago, Eiuol said:

If they were deleting results, I think you would have a point.

They’re deleting user-based results in auto-completes. And in search results they’re suppressing popular pages in favor of “authoritative” ones. It seems clear now that Google has less interest in objectively presenting information. They increasingly attempt to skew the information to reflect their own worldview. So now I have to worry about how Google is trying to fool me, in addition to not getting the type of service I’ve come to enjoy.

14 hours ago, Eiuol said:

This is not how conditioning works. That’s just a scare word. Being conditioned requires a reward (or disincentive). All you have here is Google offering search results with nothing to lure you into specific links.

The reward is the search or auto-complete result itself. You type in a word, and then you are immediately rewarded (or not) with an auto-complete phrase or a list of pages. It is absolutely conditioning. And Google knows this. Fortunately, people are not dogs in a laboratory, and they won’t tolerate being fed BS for very long.

[MisterSwig]

I’m a little obsessed with this auto-complete thing now. I tried a different search term that’s not in the Project Veritas video. I picked “joe biden groping”.

Compare Bing to Google:

I didn’t even get past the “g” before Bing’s auto-complete offered “groping children.” Now check out Google.

I typed the whole thing and got nothing. All results deleted. It stopped offering me auto-completes after “grop”.

And this is all I got at “gro”.

How come I got “grope” auto-completed after typing “gro” but not “grop”?

[MisterSwig]

Let’s do one more, as proof that Google is trying to protect Democratic candidates by manipulating search features. Compare them to Bing and Yahoo regarding the search string “elizabeth warren lie”.

 

 

Ridiculous. Yahoo and Bing deliver believable results, while Google offers nothing at all. Google is dynamiting its own product and reputation for the sake of these Democrats’ reputations.

Also, auto-complete isn’t just a suggestion. It’s an attempt to complete your phrase for you. It can therefore have the effect of shifting your attention from one thought to another. You might start with the intention of looking up the latest Democrat controversy, only to let Google redirect you to something else, because that’s what appears in the auto-complete, and you’re an easily distracted, manipulatable type of user.

[_thenelli01]

Somehow Elizabeth Warren nat …”Ural hair color” shows up and “native american” doesn’t.

[Eiuol]

9 hours ago, MisterSwig said:

. They’re being dishonest with the users who have come to expect a certain level of neutrality.

The thing I disagree with is the sinister portrayal in the video, basically manipulative propaganda through sound and word choice, hoping to add a level of evil beyond disapproval of Google’s values.

All companies (good ones, bad ones, and all the ones in between) have values, which implies political values, especially multinational corporations. Effective searching requires using a method to rank values, depending on your standard of searching. If a company is providing you with the search service, it’s up to them to pick their method of ranking values. I wouldn’t fault a company for using their own values to create a method of ranking. That includes if they use their political values.

“Suppressing” here is just another way to say “they aren’t always ranking websites according to page popularity”. Actually, this is what I would want from a search engine. I want authoritative sources first. That actually would be an objective presentation of information. But that doesn’t mean the presentation is absent of any value judgment. All information someone presents to you is skewed towards a worldview. There is a level of dishonesty when Google isn’t transparent about how they present the information, but I don’t think it rises to the level of 1984 political manipulation.

I probably should have explained it differently. Google provides reward in the broad sense that you ask for something, and they give you something. Some people might argue that capitalism is by nature bad because of this. We have a need to get stuff, as taught by capitalism, and which is made more salient through marketing. So we ask for that thing, like an iPhone, then the capitalist gives it to us. We are rewarded for our obedience to the system. This turns us into slaves of the system, treated like dogs, and reduced to commodities. The problem with this sort of argument is - is the “system” producing the initial stimulus to buy something, or is it the individuals? Or is it some combination? Classical conditioning, the kind you’re thinking of, requires a stimulus in a (relatively) controlled setting. Google is not providing a stimulus. Maybe you could describe this as a type of conditioning (it would be operant conditioning, [hitting a lever, without any apparent stimulus causing it, getting food in return] which Skinner claimed was how anyone learned anything at all), but there are problems with this just like that argument I gave against capitalism.

6 hours ago, MisterSwig said:

It’s an attempt to complete your phrase for you. It can therefore have the effect of shifting your attention from one thought to another. 

Isn’t this literally what a suggestion is?

6 hours ago, MisterSwig said:

You might start with the intention of looking up the latest Democrat controversy, only to let Google redirect you to something else

Why didn’t you look up “latest Democrat controversy”? Just finish typing it in. When I searched that, this is the first result I got: https://www.politico.com/story/2019/06/25/joe-biden-2020-democrats-1380317  I mean, I don’t think most people have such bad attention that seeing auto complete would push someone to completely change their intended behavior. And if someone is that much of a zombie, well, it is better to show them less propaganda anyway.

7 hours ago, MisterSwig said:

How come I got “grope” auto-completed after typing “gro” but not “grop”?

Because “growing up” and “group” aren’t spelled with “grop”. By the way, I didn’t get any of the results you did. Have you searched those terms before? I never have, but if you have, it would make sense. Also the auto complete on Bing is probably worse because “lieawatha” it is really just click bait. Google appears more reserved than simply listing by popularity. And of course, is probably influenced somewhat by their political values.
 

[MisterSwig]

18 hours ago, Eiuol said:

The thing I disagree with is the sinister portrayal in the video, basically manipulative propaganda through sound and word choice, hoping to add a level of evil beyond disapproval of Google’s values.

Okay, the style is overly dramatic and corny, but that’s been the case with most of his videos. I focus on the substance, mainly what the whistleblower is saying.

18 hours ago, Eiuol said:

If a company is providing you with the search service, it’s up to them to pick their method of ranking values.

They aren’t ranking them, though. They’re removing them, in the case of the auto-completes. With the page listings, it’s annoying that they bury content. I’d rather have the option to filter the results by different criteria, because I’m less interested in what Google thinks is valuable to me. I’m the one looking things up, not them. You see how it’s a service issue?

18 hours ago, Eiuol said:

All information someone presents to you is skewed towards a worldview.

I want it skewed towards my worldview. They should allow filtering like they do on YouTube video searches, by relevance, date, type, etc. They already have a News tab filter. Why can’t the All tab filter by popularity? Must both be tuned to their authoritative sources?

19 hours ago, Eiuol said:

Google is not providing a stimulus.

Keeping in mind that we’re talking about volitional humans, not dogs, the stimulus is the search page. You open your browser and see a search box. That stimulates the response to enter some words in the box. Then you are variously rewarded (or not) based on Google’s algorithm. Because we’re talking about a conditioned response here, it can be easily overcome once the user no longer gets pleasure from the reward. In fact, I’m already thinking about switching to Yahoo or Bing for my Internet searching needs.

19 hours ago, Eiuol said:

Isn’t this literally what a suggestion is?

Auto-complete is more than a mere suggestion. It’s also an interruption. It’s like when you start a question, “Have you seen my–”, and your friend interrupts, “your baseball cap?” But maybe you were going to say, “car keys.”

19 hours ago, Eiuol said:

Why didn’t you look up “latest Democrat controversy”?

Because I was testing the auto-complete. Besides, looking up “joe biden groping” is more relevant. It’s not like they stop you from typing in your phrase. That’s not the issue.

19 hours ago, Eiuol said:

I mean, I don’t think most people have such bad attention that seeing auto complete would push someone to completely change their intended behavior.

Even if you don’t get initially distracted, if your intention is to look up popular searches that include your phrase, then you’re out of luck. Google won’t let you see them. If I never used Bing, I wouldn’t have learned about “elizabeth warren lieawatha.” And maybe that information interests me.

19 hours ago, Eiuol said:

By the way, I didn’t get any of the results you did. Have you searched those terms before?

No. Prior searches appear in the list with a little clock icon next to them. Can you provide a screenshot of your result?

[Eiuol]

3 hours ago, MisterSwig said:

They’re removing them, in the case of the auto-completes

I don’t really understand what you mean by “removing”. Auto complete isn’t even made out of search results, there isn’t anything to take away that used to be there. 

Yeah, a service issue. I agree. Not a grand agenda to manipulate the populace. There are plenty of moral problems with Google, but I don’t think this is one of them. Their worldview is skewed a different direction than yours, which isn’t really a threat to your very existence. Immoral, maybe, sinister, no.

3 hours ago, MisterSwig said:

Keeping in mind that we’re talking about volitional humans, not dogs, the stimulus is the search page.

Dogs are volitional in some sense. Anyway, principles of associative learning (conditioning) are only valid if you accept that all animals, even humans, operate in part through these principles. That isn’t to say -all- learning is associative, only that people will also learn through association. The part that matters is the additional behaviors we can engage in. Breaking the association is often difficult to overcome. It also means that associative learning occurs in specific environments and contexts. It doesn’t simply happen any time there is a stimulus, and it matters where the stimulus comes from. 

It’s fair to call a search box a (behavioral) stimulus, but Google didn’t make this so. More accurately, Google has provided a potential stimulus (but probably became so after you so many search boxes and so many places), but they aren’t giving you a new stimulus that is associated. This makes it so principles of associative learning don’t really apply. If Google really wanted to condition you, they would do it differently, in the ways that actual manipulative companies do.

3 hours ago, MisterSwig said:

Even if you don’t get initially distracted, if your intention is to look up popular searches that include your phrase, then you’re out of luck.

Google isn’t the best place to do that anyway because their search algorithm isn’t based on popularity of pages. It never was, and that’s common knowledge.
 

3 hours ago, MisterSwig said:

No. Prior searches appear in the list with a little clock icon next to them. Can you provide a screenshot of your result?

I didn’t get any auto complete results after I got to “gro”. If it matters any, I’m not on mobile, so that could explain the difference. Or maybe your region. I got the same things as you from Bing and Yahoo, though.

[MisterSwig]

22 minutes ago, Eiuol said:

I don’t really understand what you mean by “removing”. Auto complete isn’t even made out of search results, there isn’t anything to take away that used to be there. 

Auto-complete, as far as I know, pulls data from a database of search strings that have been entered by users. So if there is a big event, say an earthquake in Tokyo, people will start searching for “earthquake in Tokyo.” Google’s machine learning will recognize this rather quickly as the popular search phrase. So when I start typing in “earthquake,” the auto-complete will suggest “in tokyo” because that’s what most people are looking for at that moment. I’m saying that Google, in these political situations, blocks me from seeing such relevant and popular auto-completes. They must be in the database, because Google is storing your search strings and using them for other auto-completes. But in certain situations, they suspend the feature from giving you the data, or they draw data from the special “source of truth” database.

34 minutes ago, Eiuol said:

It’s fair to call a search box a (behavioral) stimulus, but Google didn’t make this so. More accurately, Google has provided a potential stimulus (but probably became so after you so many search boxes and so many places), but they aren’t giving you a new stimulus that is associated. This makes it so principles of associative learning don’t really apply.

In operant conditioning, I believe it’s called a discriminative stimulus. I’m not saying that the conditioning part is wicked. That’s how we’ve learned to use the search box to get information and associate it with value. I’m saying that they’re now using that conditioned response to feed us an inferior product. So naturally those of us who don’t want to consume garbage will have to go elsewhere. But it’s tough, because we’ve become habituated to using Google.

57 minutes ago, Eiuol said:

I didn’t get any auto complete results after I got to “gro”. If it matters any, I’m not on mobile, so that could explain the difference. Or maybe your region.

Interesting. I haven’t tested a desktop. And I have no clue if the results are different by region. I’m in Los Angeles.

[Eiuol]

9 hours ago, MisterSwig said:

I’m saying that they’re now using that conditioned response to feed us an inferior product.

You’re right, the search box would be a discriminative stimulus. But I don’t think it would help any for consuming an inferior product. Part of the reason is that the reinforcement is pretty weak (enough so that I don’t think it should be called reinforcement to get the results of the search). It doesn’t push you to act, because you already decided to search for something. At best, auto complete would show you something that might pique your interest, but there is no particular preferable reward if you already had in mind something you are searching for. You’d have to already be inclined and interested in the auto complete results for them to have an effect. On top of that, each case of auto complete is a different case of learning in terms of association. That’s why you have lost interest over time - the auto complete results have become less valuable. In the end, conditioning just wouldn’t work.

Of course, I would still suggest using other products to anyone. To me though, to fix people using Google for their searches, you’d have to lead them to prefer more objectivity. Muckraking articles have their place (but it really rubs me the wrong way when what is presented more like propaganda than objective journalism), though the more actionable thing to do is suggest a replacement value.

[MisterSwig]

12 hours ago, Eiuol said:

You’d have to already be inclined and interested in the auto complete results for them to have an effect.

I think most people are naturally curious about the effects of their own actions. My typing causes the auto-completes to appear, and so I’m interested in what I just caused to happen. Additionally, merely seeing and reading the auto-completes causes me to think about them. So there is a basic effect they have on me, even if I don’t select them.

Another thought: if people don’t realize that the product is inferior, then the prior conditioning will remain, everything else being equal. By Google omitting only some auto-complete results, it’s harder to detect the change, because people will simply assume there were no results available. It’s like when Taco Bell removes a menu item. Only those who notice and enjoy that item will consider it an inferior menu. Others will be equally stimulated by the menu as before, because they don’t notice any negative change.

Furthermore, in some cases, the product might remain in appearance but not quality, and only those who can detect the inferior quality will experience an extinction of the conditioned response. I’m thinking of cases where Google substitutes non-user generated auto-completes for user-based ones. If you don’t want the non-user ones, and Google doesn’t label the difference, then you’ll value the product less, but only if you know that Google does this. It’s like restaurants labelling the difference between artificial and real ice cream. If they don’t tell you that their ice cream is artificial, you might assume it’s real if it looks and tastes like real ice cream. And you would remain conditioned to order it, under that false belief.

[Eiuol]

11 hours ago, MisterSwig said:

I think most people are naturally curious about the effects of their own actions

This is why conditioning doesn’t much apply. Exploration through curiosity that produces learning isn’t conditioning; it isn’t associative learning. If your explanation involves curiosity and contemplation, you’ve already ruled out conditioning as a good explanation for the behavior. You could argue that everything boils down to conditioning, but I really doubt you’re trying to say that anyway.
 

11 hours ago, MisterSwig said:

if people don’t realize that the product is inferior, then the prior conditioning will remain, everything else being equal.

If people can’t tell the difference, then in what way is it even an inferior value? If you’re concerned that people are mindless zombies, which is a valid concern, it’s not because they were conditioned to be so. Google is not helping, but they aren’t causing it. I’m much quicker to blame the actual content that’s out there than gatekeeping. 

And extinction of what? When you talk about extinction, responses, stimuli, and all that stuff, constant generalization isn’t really going to work here. Each search is very different. Each search, if it were a controlled experiment, would be a new instance of learning. “Using Google” is way too broad. Generalization of that sort requires abstraction.

[MisterSwig]

32 minutes ago, Eiuol said:

If people can’t tell the difference, then in what way is it even an inferior value?

Because values are objective. It’s not a big deal with real vs. fake ice cream, or with auto-completes. But let’s imagine you buy a million-dollar painting. You think it’s a real Van Gogh, but it’s actually a fake. All your life, you look at it, enjoy it, and believe you have invested wisely in a precious piece of art. Then you die, and your children find out it’s a forgery. It’s worthless. They have no inheritance. Now you’re remembered as the sucker who gave all his money to criminals. This is why people don’t like being fooled or tricked. They want to avoid getting less than what they’re promised. If I’m going to use Google and generate ad views for them, I expect the user-based auto-completes that they offer and describe in their guidelines. They don’t say anything about some “source of truth” database or removing auto-completes to protect political candidates.

[Eiuol]

9 hours ago, MisterSwig said:

Because values are objective.

“Values are objective” means they are agent relative, but relate to the nature of reality. A fake van Gogh is a value qua value. There is no fact embedded in a value that exists apart from your awareness, because value is a psychological thing. You may be tricked about facts when people hide what you care about, but you would value things equally if you see no difference between them at all. Of course, you might want to tell people that their values are misplaced, that Google is doing things they should notice for the sake of their life, there just isn’t any inferiority until they recognize the difference.

[MisterSwig]

22 hours ago, Eiuol said:

There is no fact embedded in a value that exists apart from your awareness, because value is a psychological thing.

I don’t know what you mean by this. Values exist relative to living things, not psychological things. A value is a value to us whether we are aware of it or not. The same goes for disvalues that we mistake as values. If I eat poison thinking it’s ambrosia, it’s not a value to me simply because I’ve mistaken it as such. It’s objectively bad for me. In this case, the negative effect of my mistake might take only a few minutes to occur, as I become ill or possibly die. In the case of the fake painting, the primary negative effect (financial loss) might be unrealized for years. Then there is the secondary effect on my reputation and/or self-esteem. Instead of a wise investor in art, I’m now an ignorant victim of fraud.

[MisterSwig]

So Google rationalizes removing auto-completes by claiming that they are protecting individuals from being associated with “sensitive and disparaging terms.” I guess this means that “Bill Clinton” is now considered such a term.

Exhibit 1

Exhibit 2

That sure is one smart algorithm!

And just in case you’re wondering…

Exhibit 3

Exhibit 4

[Eiuol]

On 6/29/2019 at 10:08 AM, MisterSwig said:

Values exist relative to living things, not psychological things.

The value is something you hold psychologically. That’s why we would say you need a code of values, rather than finding true values. At least by Oism, values are to be selected and decided upon. I remember Peikoff explaining something about this. As far as I remember, he was saying we begin with inducing the concept of value from all kinds of observations of people desiring or wanting things, then acting to gain those things. So we can talk about values on the one hand as the general psychological phenomena, and we can also talk about values according to the Objectivist standard and deny that something is of value.

But we aren’t talking about a narrow understanding of value specific to our code of values. If people prefer nonobjective information, that is still there value, and they will seek it out, and they should have the freedom to do so. The important question I think is not so much what Google is doing (because people actually value what they provide, even if it is not objective information), but how to get people to care about objective information. Google is not the cause, it’s just a consequence of people caring less about objective information anymore.

Also, part of the reason lying and fraud is bad is because with false information, you can’t fairly or accurately determine your values according to how reality is. But either way, you still wanted the fake painting, so you valued the fake painting.