Thread by @AriSchulman

In reply to Trump admin "taking a look" at regulating search, Google says they bias results by relevance, not politics. The problem is that Google has spent years redefining good information in political terms. You must read @adamjwhitedc in @tnajournal: thenewatlantis.com/publications/g…

Here's the WaPo piece, which isn't helping the paranoia surrounding this question by only citing the concerns of POTUS, Diamond & Silk, etc. washingtonpost.com/news/morning-m…

Here are some of the examples. While Google had for several years targeted payday lenders for spamming practices, in 2016 they banned payday-loan ads outright -- not for abusing search policies but based on their political reasoning about socially beneficial policy.

The timing was very close to the Obama administration's own targeting of payday lenders.

Three months ago, Google also banned ads for bail-bond services.

The timing of these decisions is significant. It suggests that Google wasn't making decisions in a vacuum but was doing so in concert with broader political and lobbying initiatives, underscoring the... not exactly apolitical nature of these decisions.

Also, of course, Google reopened its censored search in China, bowing to the demands of the Chinese regime. Google execs, when they previously allowed censored search in China, decided on it using "moral metrics" and an "evil scale."

Ditto (on a much lesser scale) removing neo-Nazi pages in France and Germany.

Also striking is Google's recent ban of ads related to the Irish abortion referendum, again offering a supposedly neutral rationale that was obviously chosen ad hoc -- and again clearly spurred by political pressure.

There's a great deal more in the piece on how Google has become a powerful lobbying force over the last decade, and has long thought itself as having a role in what information users should find. The "we're not doing politics" pretense can't be sustained.
thenewatlantis.com/publications/g…

Google in 2009: "we do not remove a page from our search results simply because its content is unpopular."
In 2017: "offensive or clearly misleading content ... is not what people are looking for," and "We've adjusted our signals to help surface more authoritative pages"

Finally is this remarkable comment from Alphabet chairman Eric Schmidt, advocating de-ranking pages based on Google's judgment of truth and falsehood:

It's possible not to take the POTUS line that Google has a downrankGOP() procedure in their algorithm and still take the collection of Google's words and actions at face value and believe that they increasingly see political intervention in search as a feature, not a bug.

Someone unrolled this thread: threadreaderapp.com/thread/1034525…

It's really hard to get too worked up about Google removing/downranking Holocaust denial pages, I know. But go look again at the shift between their 2009 and 2017 statements.

Google has plainly shifted policy, from avowedly not altering search based on offensiveness to avowedly doing so. But they say this is *just about relevance,* that offensive/misleading content "is not what people are looking for." This rationale is the most troubling part.

It's one thing for Google to say "we can't abide hate / fake news" and own up to how downraking conflicts w being a neutral gateway. It's another to say "we're just giving people what they actually want" when the reason these results show up is that people *do* want fake news.

It's hard to tell whether Google understands how disingenuous it's being, and whether it's worse if they buy their own line. But this rationale and its potential capaciousness is the part of this story that provides non-crazy reason to wonder how far Google might go with this.

Bump.