Thread by @doctorow

Google's search quality has been in steady decline for years, and Google assures us that they're working on it, though the most visible effort is replacing links to webpages with lengthy, florid paragraphs written by a confident habitual liar chatbot:

pluralistic.net/2023/02/16/twe…

1/

If you'd like an essay-formatted version of this thread to read or share, here's a link to it on pluralistic.net, my surveillance-free, ad-free, tracker-free blog:

pluralistic.net/2023/06/15/pai…

2/

The internet is increasingly full of garbage, much of it written by other confident habitual liar chatbots, which are now extruding plausible sentences at enormous scale.

3/

Future confident habitual liar chatbots will be trained on the output of these confident liar chatbots, producing @jathansadowski's #HabsburgAI:

https://twitter.com/jathansadowski/status/1625245803211272194

4/

But the declining quality of Google Search isn't merely a function of chatbot overload. For many years, Google's local business listings have been *terrible*.

5/

Anyone who's tried to find a handyman, a locksmith, an emergency tow, or other small businessperson has discovered that Google is worse than useless for this. Try to search for that locksmith on the corner that you pass every day?

6/

You won't find them - but you *will* find a fake locksmith service that will dispatch an unqualified, fumble-fingered guy with a drill and a knockoff lock.

7/

They'll drill out your lock, replace it with one made of bubblegum and spit, and charge you 400% the going rate (and then maybe come back to rob you):

nytimes.com/2016/01/31/bus…

8/

Google is clearly losing the fraud/spam wars, which is pretty awful, given that they have spent billions to put every other search engine out of business.

9/

They spend $45b every year to secure exclusivity deals that prevent people from discovering or using rivals - that's like buying a whole Twitter every year, just so they don't have to compete:

thebignewsletter.com/p/how-a-google…

10/

But there's an even worse form of fraudulent listing on Google, one they *could* do something about, but choose not to: ad-fraud. For all the energy thrown into "dark SEO" to trick Google into putting shitty, scammy sites atop of the listings, there's a far simpler method.

11/

All you need to do is pay Google - buy an ad, and your obviously fraudulent site will be right there, at the top of the search results.

12/

There are so many top searches that go to fraud or malware sites. Tech support is a favorite. It's not uncommon to search for tech support *for Google products* and be served a fake tech-support website.

13/

These allow scammers to trick you into installing a remote-access trojan and then steal everything you have, and/or take blackmail photos of you with your webcam:

bleepingcomputer.com/news/security/…

14/

This is true even when Google has a trivial means of reliably detecting fraud. Take the restaurant monster-in-the-middle scam: a scammer clones the menu of a restaurant, marking up their prices by 15%, and then buys the top ad slot for searches for that restaurant.

15/

Search for the restaurant, click the top link, and land on a lookalike site. The scammer collects your order, bills your card, then places the same order, in your name, with the restaurant.

16/

The thing is, Google runs these ads even for restaurants that are verified merchants - Google mails the restaurant a postcard with a unique number on it, and the restaurant owner keys that number in to verify that they are who they say they are.

17/

It would not be hard for Google to check whether an ad for a business matches one of its verified merchants, and, if so, whether the email address is a different one from the verified one on file.

18/

If so, Google could just email the verified address with a "Please confirm that you're trying to buy an ad for a website other than the one we have on file" message:

pluralistic.net/2023/02/24/pas…

19/

Google doesn't do this. Instead, they accept - and make a fortune from - paid disinformation, across every category.

20/

But not all categories of paid disinformation are equally bad: it's one thing to pay a 15% surcharge on a takeout meal, but there's a whole universe of paid *medical* disinformation that Google knows about and has an official policy of tolerating.

21/

This paid medical disinformation comes from "#CrisisPregnancyCenters": these are fake #AbortionClinics that raise huge sums from religious fanatics to buy ads that show up for people seeking information about procuring an abortion.

22/

If they are duped by one of these ads, they are directed to a Big Con-style storefront staffed by people who pretend that they perform abortions, but who bombard their marks with falsehoods about health complications.

23/

These con artists try to trick their marks into consenting to sexual assault - a transvaginal ultrasound. This is a prelude to another fraud, in which the "sporadic electrical impulses" generated by an early fetal structure is a "heartbeat."

24/

(Early fetuses do not have hearts, so they cannot produce heartbeats):

nbcnews.com/health/womens-…

25/

If the victim still insists on getting an abortion, the fraudsters will use deceptive tactics to draw out the process until they run out the clock for a legal abortion, procuring a forced birth through deceit.

26/

It is hard to imagine a less ethical course of conduct. Google's policy of accepting "crisis pregnancy center" ads is the moral equivalent of taking money from fake oncologists who counsel people with cancer to forego chemotherapy in favor of juice-cleanses.

27/

There is no ambiguity here: the *purpose* of a "crisis prengancy center" is to deceive people seeking abortions into thinking they are dealing with an abortion clinic.

28/

They then further deceive them into foregoing the abortion, by means of lies, sexually invasive and unnecessary medical procedures, and delaying tactics.

29/

Now, a new report from the @CCDHate finds that Google made $10m last year on ads from "crisis pregnancy centers":

wired.com/story/google-m…

30/

Many of these "crisis pregnancy centers" are also registered 501(c)3 charities, which makes them eligible for Google's ad grants, which provide free ads to nonprofits.

31/

Marketers who cater to "crisis pregnancy center" advertise that they can help their clients qualify for these grants. In 2019, Google was caught giving tens of thousands of dollars' worth of free ads to "crisis pregnancy centers":

theguardian.com/technology/201…

32/

The keywords "crisis pregnancy centers" bid up include "Planned Parenthood" - meaning that if Planned Parenthood clinics wanting to appear at the top of the search for "planned parenthood," must to outbid the fraudsters seeking to deceive Planned Parenthood patients.

33/

Google has an official policy of requiring customers that pay for ads matching abortion-related search terms to label their ads to state whether or not they provide abortions, but the report documents failures to enforce this policy.

34/

The labels themselves are confusing: for example, abortion travel funds have to be labeled as "not providing abortions."

35/

Google isn't afraid to ban whole categories of advertising: for example, Google has banned #PlanC, a nonprofit that provides information about medication abortions. The company erroneously classes Plan C as an "unauthorized pharmacy."

36/

But Google continues to offer paid disinformation on behalf of #ForcedBirth groups that claim there is such a thing as "#AbortionReversal" (there isn't - but the "abortion reversal" drug cocktail is potentially lethal).

37/

This is inexcusable, but it's not unique - and it's not even that profitable. $10m is a drop in the bucket for a company like Google. When you're lighting $45b/year on fire just to prevent competition, $10m is chump change.

38/

A better way to understand Google's relationship to paid disinformation can be found by studying #Facebook's own paid disinformation problem.

39/

Facebook has a well-documented problem with paid political disinformation - unambiguous, illegal materials, like paid notices advising people to remember to vote on November 6th (when election day falls on November 5th).

40/

The company eventually promised to put political ads in a repository where they could be inspected by all parties to track its progress in blocking paid disinformation.

41/

Facebook did a *terrible* job at this, with huge slices of its political ads never landing in its transparency portal.

42/

We know this because independent researchers at NYU's engineering school built an independent, crowdsourced tracker called #AdObserver, which scraped all the ads volunteers saw and uploaded them to a portal called #AdObservatory.

43/

Facebook viciously attacked the NYU project, falsely smearing it as a privacy risk (the plugin was open source and was independently audited by #Mozilla researchers, who confirmed that it didn't collect any personal information).

44/

When that didn't work, they sent a stream of legal threats, claiming that NYU was trafficking in a "circumvention device" as defined by #Section1201 of the #DigitalMillenniumCopyrightAct, a felony carrying a five-year prison sentence and a $500k fine - for a first offense.

45/

Eventually, NYU folded the project. Facebook, meanwhile, has fired or reassigned most of the staff who work on political ad transparency:

pluralistic.net/2021/08/06/get…

46/

What are we to make of this? Facebook claims that it doesn't need or want political ad revenue, which are a drop in the bucket and cause all kinds of headaches.

47/

That's likely true - but Facebook's aversion to blocking political ads doesn't extend to spending a lot of money to keep paid political disinfo off the platform.

48/

The company *could* turn up the sensitivity on its blocking algorithm, which would generate more false positives, in which nonpolitical ads are misidentified and have to be reviewed by humans.

49/

This is expensive, and it's an expense Facebook can avoid if it can suppress information about its failures to block paid political disinformation. It's cheaper to silence critics than it is to address their criticism.

50/

I don't think Google gives a shit about the $10m it gets from fake abortion clinics. I think Google believes the PR trouble it would get into for blocking them - and the expense it would incur in trying to catch and block fake abortion clinic ads - are real liabilities.

51/

In other words, it's not about the $10m it would lose by blocking the ads - Google wants to avoid the political heat it would take from forced birth fanatics and cost of the human reviewers who would have to double-check rejected ads.

52/

In other words, Google doesn't abet fraudulent abortion clinics because they share the depraved sadism of the people who run these clinics. Rather, Google teams up with these sadists out of cowardice and greed.

53/

Image:
Flying Logos (modified)
commons.wikimedia.org/wiki/File:Over…

CC BY-SA 4.0
creativecommons.org/licenses/by-sa…

eof/