For those of you who have not seen Matt Taibbi’s piece on Google’s threat to demonetize us over flawed, even nonsensical AI-generated charges against a handful of posts since 2018, please proceed immediately and give it a read. We are very grateful for his deft skewering of the Google spreadsheet which was the only explanation of its charges, and his trying to fathom the logic, such as it was, behind Google demonetizing the designated posts. Recall that Google was threatening to demonetize the site entirely if we did more of the same….when the accusations on their face (save possibly one) made no sense.
Taibbi nailed one of the key reasons why the Google sanctions were so off base: “…this is a common feature of moderation machines; they can’t distinguish between advocacy and criticism.” And he also cited examples where articles in highly respected medical journals which raised doubts about Covid vaccine performance triggered warnings like [HARMFUL_HEALTH_CLAIMS, ANTI_VACCINATION, HATEFUL_CONTENT].
Again, we are extremely appreciative of Taibbi getting on this case so quickly and producing a compact, easy-to-digest (important for generalist readers!), yet very effective piece, and then at remarkable speed too.
So please, do your part by circulating Taibbi’s piece on social media, such as Twitter, Facebook, LinkedIn, and relevant sites like YCombinator and appropriate forums on Reddit, and posts or sites that focus on free speech issues. Things like this never happen in onesies. As we recounted earlier, the one publisher we knew who got a lesser roughing up by Google (no site-wide demonetization threat) did get Google to back down on some charges, but he capitulated on the rest. And critically, he was insistent I not identify him, which in context I took to mean that he saw the Google demerits as potentially reputation-damaging. So there may already be a lot more of this sort of Google content-censoring than has been made public.
Again, thank Taibbi for his help!
65 comments
Comments are closed.
And along with your donation to NC maybe throw a few dollars Taibbis way… He’s gotta eat too and this stuff isn’t going to find its way onto the front pages of WAPO or the NYT.
Just my two cents.
Yes, excellent suggestion!
$5 a month gets you a subscription to Taibbi’s “Racket News” which features all his writing plus his weekly conversation with Walter Kirn. Its a great deal!
second that. Matt is proving to be a national treasure even if he sometimes gets hold of the wrong end of the stick (Battacharya)—an occupational hazard for free speech absolutists.
Or perhaps he is willing to follow the story wherever it leads. Freedom of speech means freedom of speech that pokes your neuroses.
Subscribed to Taibbi…
…finally!
Had to get Doc Hudson 1st!
Well, Naked Capitalism may not have violated the letter of Google’s capricious rules, but it certainly violated their unstated spirit of penalizing any challenge to the status quo. Thus you cannot blame AI for malfunctioning, it served its master’s intent quite faithfully. Which goes to show the risk is not rogue AI, but rogue humans using AI to amplify their misdeeds.
That was my thought also. It’s absolutely working as intended.
Which just goes to note that it is neither artificial nor intelligent.
I like the article but the part where the person was talking about Trump election denial and no we dint say Biden dint win so please don’t sensor us ,not in these words but that I got. This is bad news because this proof people have kneel to the democrat party and start to accept censorship. Coming November this will go in high gear. I bet many of these people who gave government so much leeway with censorship and Trump win 2024 they will regret not pushing against democrat party harder against Internet censorship
We said no such thing. We are saying how the algo is assigning punishments.
Sorry, I meant in taibbi article
Given the Streisand effect, this may result in a decent reader bump. Hacker news has picked it up.
The #fediverse don’t care about google. That crazy (family blog) honey badger doesn’t give a (family blog) about corporate power.
https://www.youtube.com/watch?v=4r7wHMg5Yjg
No, I had not seen that yet, thanks for informing us. Matt Taibbi is one of the best independent journalists out there.
His support for Julian Assange is admirable as well.
https://scheerpost.com/2023/06/25/why-julian-assange-must-be-freed/
AI Censorship Targets People Who Read Primary Sources To Fact-Check The News
https://thefederalist.com/2024/03/07/ai-censorship-targets-people-who-read-primary-sources-to-fact-check-the-news/
Thanks a million, Mary. My hubby and I often debate the reach of the 1st amendment. He was pushing how the NYT was mocking people who “did their own research”. He finally gave up when we talked about doing research before buying a car. He’s back at it now with the comment by supreme court justice Brown Jackson that free speech “hamstrings the government”.
It is a bad trend to emphasize the inaccuracy of AI. As algorithms become more accurate, censorship will become more sophisticated and the situation will get worse. The problem is not the quality of censorship, but the act of censorship itself. Google has no authority to determine what constitutes a “democratic process” or the accuracy of information about vaccinations.
Au Contraire. The 1st Amendment very much gives them the authority. It’s up to the user to determine whether what they are viewing is worth a (family blog). The alternative is to go somewhere else — which leads to the anti-trust arguments.
The better argument is Google as a “common carrier” has a *duty* not to injure the public by intentional suppression of search results we need to be a functioning society. That analysis steps out of the constitutional law world and into the tort world.
And, once again the GitHub #fediverse community comes swinging to the rescue wit the Vivaladi web browser: https://en.wikipedia.org/wiki/Vivaldi_(web_browser)
All the techies, who understand the stakes, are moving away from the walled garden corporate approach. Getting laypersons to leave their comfy walled gardens is the challenge. We have the power to make corporate domination of the interwebs a thing of the past through the #fediverse.
The choice is this: Endure the switching costs, or continue to sanction the slide into 1984 by implictly sanctioning corporate malfeasance by not switching.
And: We can look to the revelations of Ed Snowden and Julian Assange/Wikileaks, for the connections between the BigTech monopolists and US “intelligence”. Is it dystopian enough yet?
No wonder the UK/US is holding Assange in solitary at a Max Sec. Belmarsh Prison; and Snowden had to flee to Russia. There is a ton of info on this, here is just a couple of links
https://www.wired.com/2013/10/nsa-hacked-yahoo-google-cables/
https://wikileaks.org/google-is-not-what-it-seems/
It seems to me, that the USA has been colonized by a foreign power……. A power with it’s own, distinct, set of values, laws and history, …….very much as India was colonized by the English..
British Colonization of India
The British had a clear goal: to westernize and Christianize India, using all possible means.
“……But, besides monetary prospects, another significant reason/motivation for all European colonisations in different parts of the world is almost always ignored. It was an inherent ideology of “western supremacy”. This idea is often referred to as “the white man’s burden”: a self-proclaimed responsibility of the west to subjugate and civilize any other person, who does not fall under the category of the “west”.
Accordingly, the Europeans, who “discovered” India had both intentions in mind. But, it was the British, who were the most successful among them. They not only succeeded in controlling most of the Indian Territory, but also successfully colonized various aspects of the Indian society like culture, politics, economy and education. (Indian)”
“….It should also be noted that the Indian economy, which contributed 23% of the world economy during the 17th-18th centuries was down to around 4% in 1947. Such was the loot conducted by the British for most of their period in India. This, however, does not capture the exact picture of the sufferings of the then people.”
https://indiafacts.org/british-colonization-india/
Sorry I don’t understand your reply. Colonized by whom? The US is the (declining) global hegemon. The US has military bases across the world. We speak English for a reason.
Imperialism, war, and slaughter are as “American as apple pie”. (And as British as Marmite and Fish n Chips)
It might be comforting or convenient to blame somebody else, but the US oligarchy runs the show.
Vivaldi? Really? Went to your link and found it interesting but even Wipepedia (not reliable) mentions “users can install extensions directly from the Chrome Web Store.” Chrome is very poor regarding surveillance. So do you have more info about Vivaldi and how functional it is while remaining under the umbrella of Gaggle? I’m happy so far with Brave since it avoids routing through Gaggle. See Joe Rogan’s episode #1768 with Robert Epstein – he chews up Gaggle.
Yup. You’re right. I went down a rabbit hole based on incomplete info. And for the record: I use Brave.
A handful of us switching browsers or search engines doesn’t help a bit when the entire internet runs through Google’s servers. Watch all the intermediate sites hit when you click on any web address. Even government servers route you through Google World. There’s a reason ALL web travel and search sucks now, and its name is Google.
Vivaldi is made by old Opera people. It does use Chromium rendering engine as do all other popular browsers except Firefox. Chromium and Apple’s Safari is based on Webkit. We used to have more rendering engines in the past – MS internet Explorer, old Opera’s Presto, Firefox Gecko. Now we have two and almost total Webkit monopoly, the now de facto private standard for WWW.
Vivaldi recreates old Opera functionality (mail client, now calendar, many customization options). It has integrated Mastodon client now as well, perhaps this is why it was mentioned.
I switched to Vivaldi after Firefox started crashing regularly and Chrome kept insisting I log in. Privacy Badger is the only extension I’ve installed. After some normal getting used to, I find the left side tool bar very intuitive and the especially “reading list” feature is a good match for the way I navigate the Internet.
Mail client is excellent and works on totally different paradigm than Outlook or similar. You have contacts added automatically with every mail that you respond to. You can filter contacts in real time by typing in string of text (ex. a domain). When you click on a contact only correspondence with that contact is filtered in real time. Then you type in keyword (say invoice) and you have further filter in real time from that contact. Super fast, no setting up rules to move mail to folders needed at all, you just type in keywords.
Old Opera had lots of ingenious stuff like that, it pioneered tabs and most of other stuff now taken for granted they had as first.
That is a big assumption that AI algorithms will become more accurate. What does it mean anyway. Even if humans were doing it, which is the only measure of accuracy against which AI can be compared, it would still be capricious.
Arbitrariness is certainly a feature not a bug. It instills fear and subservience and ensures sites give a wide margin in their self censorship.
That’s like, putting JDAM kits on iron bombs to industrialize genocide; by carpet bombing terrified victims, crowded into ancient tenements? It’s putting on a show of slaughtering a journalist here, “jihadi” community advocate, there & a hospital or mosque to make a point? Google’s SEOing of annoying, contradictory or inconvenient FACT down it’s memory hole, now includes efforts to shop, find product reviews, dates, assess prices, evaluate accuracy or type (as spell-checker picks PMC appropriated concepts & assigns them entirely new meanings. Try, writing colloquial English, to a less than affluent friend/ coworker. Google simply inserts Ofay yuppie euphemism?)
Why do you assume algos will become more accurate? I would say that the well-documented process of enshittification (Doctorow) will make them worse, not better. That has certainly happened with search.
Ian Welsh has a pretty good explainer on it right now that seems to mirror your point. It seems quite clear that, no, we can’t expect them to get “better.” We can only anticipate that they, and the society being slowly conditioned to their use, will get “blander.” Everything on the internet distilled down to a handful of content flavors: white bread, vanilla ice cream, and saltine cracker to name a few.
This, due to what I think you once characterized as coprophagia.
Instead of accurate or the contrary or may be the term here is biased. It is quite probable that AI will, with time, become more and more biased towards the interests of the PMC. This post is good example of such thing happening.
I am slowly coming around to Liberal Fascism as a category label. I resisted at first because I know people will it as bonkers, a contradiction in terms. Otoh, it has been the liberals who over that last decade led the push to fascism in the banality-of-evil restructuring of American politics. Plus I am sick of listening to Glenn Greenwald referring to the people deprecating and pushing to abrogate our most celebrated liberal text, The Bill of Rights, as liberals. So I’m now calling those people the New Liberal Fascists and myself as a woolly-minded liberal anarchist communist.
Sorry if this seem OT but I was listening to Taibbi and Greenwald talk about the Supremes/NTY and what they call liberals pushing for merged state/corporate surveillance, censorship, punishment of wrongthink while I was driving the car this morning. No! We should to reclaim the values of the word liberal that we like and the word itself from those PMC fascists. Hence I think New Liberal Fascists. What do you think?
In the modern, US sense of the word (not British) “liberal’ now means anti-labor, pro-oligarchy, right-wing, authoritarian warmonger with a rainbow bumper sticker and a BLM lawn sign for virtue-signalling smokescreen.
Chris Hedges wrote a whole book: Death Of The Liberal Class.
That’s the redefinition I’m not prepared to accept. There are aspects of liberal traditions that I like and would like to keep and keep the name for them. Freedom of expression, free press, privacy, etc. It’s time to call the New Liberal Fascists what they are: fascists.
When you hear liberal used to describe the Democratic Party, think neoliberal. (Classical) Liberalism vs. Neoliberal Fascism.
When I hear the term the New Left, I associate it with the New Left and the Free Speech Movement of the 1960s. Things have changed since then, haven’t they?
To steal JC Medlock’s fire: That’s right.
Liberalism seems always to have its’ blind spots; freedoms applying selectively may be foundational to the tradition. Consider John Locke, one of its intellectual fathers: charter member of, and third largest shareholder in the Royal Africa Company. He is just one of many examples we might name of ‘liberal’ slave-owners, strike-breakers, sexists, &c., &c. A consistent and progressive liberalism would look a lot like anarchism. Not a lot of those near the mainstream.
Two things I associate with liberalism: Cartesian, and other, dualism, and hypocrisy. The rest changes with the times, or, probably more accurately, with the biases of the ruling class of the day.
(I am assuming that the ideas on paper are to be judged by how they influence action; there are other ways to judge ideas, of course. I just think it’s the most relevant. I like to read the ideas, ask “for whom”, and then cross-reference with history books.)
As a result, I’d say that there is nothing particularly illiberal about censorship, using state power to crush dissidents and enemies and the like. Better, in my view, to reject liberalism and study alternatives.
Well said, Vicky. Liberalism is a pernicious doctrine which has led directly to the current destructive atomisation of society. The only good thing that can be said about it is that it is fast forwarding the end of the US Empire, and the final dissolution of the British one.
Liberal is a person fed well enough to afford mistaking symbols for material reality.
I like the old union definition: ‘A liberal is someone who is on your side (or is it at your side?) until the police draw their truncheons.’
And then the locus classicus is the 1966 Phil Ochs song — often mentioned here — ‘Love me, I’m a liberal’.
.Tom: I am slowly coming around to Liberal Fascism as a category label. I resisted at first because I know people will it as bonkers, a contradiction in terms.
Neoliberalism in its valorization of the Market — and therefore of corporate power — as the Supreme Decider is by definition fascist.
Yes, strains of liberalism deriving from the original British tradition of liberalism exist that are not fascist, as JonnyJames above points out.
Still, liberalism in its neoliberal formulation — and neoliberalism does derive from liberalism, if you go back and look at what Jeremy Bentham and similar 19th century types wrote — is always fascistic.
Ah. I see JBird4049 making essentially the same point
I call them “Notsies” because they insist they are ‘not [Family-bloggers]’. Their ‘proof’ is that they are ‘100% patriots’ in the tradition that liberated the slaves and then Auschwitz.
The technical terms “imperialist” or “colonial” are too respectful and too often read as praise, while ‘fascist’ is weak tea, watered down to the threat level of a Franco, a Mussolini, or a Macron.
I’ve switched to calling them neolibcons on social media because now that the neocons are all Biden Democrats, I fail to see any distinction between the two camps.
Read about Inverted Totalitarianism in the late Sheldon Wollin’s 2010 book/long-form essay Democracy Inc. to understand how the Clintonite dismantlement of the New Deal hijacked “liberalism” into the obsequious service of Our Billionaire Overlords.
and aside: Chris Hedges did a series of interviews with Sheldon Wollin. The interviews were still available on utube last time I looked.
AI: Another “innovation” that is proving to be Defective By Design.
First, I’m sorry for the distress all this is causing to this website directly and the implications for others that are not of the mainstream bent. Yet I wonder how censorship will get worse in a future when this kind of chilling problem will seem quaint.
Otherwise, having read the comments to Taibbi’s fine article, I just gotta note:
Wow. Some people just can’t forget getting “spanked” by Yves in the comments even years after the incident. It’s like a fetish.
I feel like I want to say something rude in the future just for the experience!
Hats off to you two for what you do. Love you or hate you, boy oh boy, do people remember you.
That’s a commendation where I come from.
I read through the comments as well. It’s fascinating to see stiff denialism in regards to covid delivered on the same breath as outrage over censorship. Connect the lines people.
The traditional left has lost its way because many progressives are now willing to say just about anything to protect their financial nest. Feminist Naomi Wolf now a regular of Steve Bannon must have concluded “conspiracy” is much easier click bait than discussing US Empire. I’ve recently concluded Thom Hartman appears to be going down a similar path. Support Matt Taibbi for no other reason than his recent reflection on his Twitter file encounter. He basically feels Elon Musk is full of it.
I’m of an opinion that it’s their belief system, not their money, that’s being protected at all cost, even the cost of reason.
Money has become their belief system. They’re simply peddlers of a failed narrative.
It’s the neoliberal belief system. The Market is the most important thing. Profit proves the rightness of anything. See this Mirowski essay in American Affairs Journal. I think neoliberal “philosophy” is dressed up greed and/or insanity. The 3rd Way Dems, the Clinton/Obama Dems believe in it completely and have done so for the last 35 years at least., imo. They believe this. Greed is good, there is no society, the Market is the true measure of all worth, money is the true measure of all value(s). The govt must be handmaiden to the Market, not it’s regulator.
Neoliberalism: The Movement That Dare Not Speak Its Name
https://americanaffairsjournal.org/2018/02/neoliberalism-movement-dare-not-speak-name/
My comment was in regard to progressives in media, not politicians. Reagan gets the blame for Neoliberalism but it was Bill Clinton, Robert Rubin and Larry Summers that made it happen. And never forget its Hillary that took the baton from the Bush Neocons. Hillary and Victoria Nuland were tight when she was Secretary of State. Joe Biden and Hillary were the parents of the Maidan Revolution in Ukraine. But regarding the progressive commentariat it’s increasingly about protecting their market share.
I won’t discount that idea, particularly amongst the “decision makers” amongst them. Yet, contemplate “party loyalty”, “voting against ones interests”, etc, and I, for one, can see that it’s not just money that drives blind allegiance.
I agree with you that at it’s not all about money, not exactly. Imo, it’s an effort to create a new machine to guide human behavior, human destiny, rather like the tech bros dream of the singularity. It’s a political movement that insists we abdicat fundamental human responsibilities, turning over those responsibilities to machines or the Market or some other reified object of human creation, imo.
Yet even then, the powerful don’t abdicate their reason, they keep stepping in to nudge the algorithms , for example, to come up with the answers they want while telling the little people its all impartial and fair, the algo said so, etc. Wolves in sheeps’ clothing?
Hmmm, this is turning into a rant. I’ll stop now. / ;)
OK, one more. (Sorry) / ;)
We’re entering The Age of the Digital Inquisition. Banish the heretics. / heh
IMO, it’s more and more related to identity politics and less and less about money as you descend through the economic strata.
Lenny Bruce used to descry “nice-nice” bourgeois ofay suburban ‘Murika, the way we’d avoided “straight people,” as the creatures who’d installed Nixon, invaded Cambodia, ratcheted US ever further into a Court mandated Oligarchic kleptocracy. As “rich hippies” morphed into yuppies, gentrifying blue collar outa red-lined homes. As counterculture just provided more FIRE Sector churn, to flip homes, offshore industrial jobs & wrangle US into rentier… ah, nevermind! Our culture, you’d appropriated was evanescent as a cloud of sinsemilla? White-flight suburbanite consumerism was the only ‘Murika available. Cycling to co-ops & CSA, with recycled containers & organic community gardens was a PHASE! An inappropriate artifact of unresolved childhood trauma, miscegenation & substance abuse.
Taibbi wrote: “Previously, a human being was sometimes involved in reviewing the final placement of sites in certain buckets, or problems might at least have been prompted by human complaints. Now, from flagging all the way through to the inevitable “Hi! You’ve been selected for income termination!” form letter, the new review process can be conducted without any human involvement at all.”
Great example of the new template for “automated media” that de-humanizes every step in the life cycle of operational data: data production, collection, analysis, and even response. The response in this instance is “Hi! You are going to be demonetised!” Note that the automation here models the logic of the military drone, which automates every step of the information collection, analysis and response process on the “battlefield.”
The entire world is now the battlefield.
All the ideological “movements” described in the comments here have nothing to do with anything. Traditional notions of “left” and “right” are to inaccurate to even talk about them. Fortunately, there is going alliance between those of us not in power. Who is in power and not is incredibly easy to see. There is an official Narrative promoted by the State (which includes Google, Big Medicine, the National Security State, the big corporate networks, movie studios, financial institutions, and so on) and anything opposed to aspects of that Narrative is becoming forbidden, those who have contrary arguments (particularly if they are grounded in logic, reason, and real science) are moving into the disinformation category and is viewed by the State as a threat and cannot be allowed. This all seems to be inexorable.
NC has always tried to connect intelligent dialog to those in power. Yves, perhaps, thinks that people in power are willing to look at reality with openness (many may well have been her classmates). But that’s not how power works at the top. There is a “line” that all members of the Uniparty must take just as the old communistas had to follow Moscow’s line. The powerful use Google and AI to control the population in the same way the imperialists in the National Security State with to control and rule the entire globe. This is all baked in the cake. As long as we continue to accept the hyper-materialism of our public structures, raw power will always dominate our political system. NC will have to accept this at some point or cease to exist in this format.
I don’t accept your proposition and neither do many others. The fact that Google is implementing this aspect of “narrative control” so badly says it does not care. It would invest resources if it did.
As Taibbi and Glenn Greenwald said on a recent System Update, the entire “narrative” is wrapped around bashing anything perceived to help Trump. What happens if Trump wins?
Once the Gooble Monopoly has been drawn and quartered, may we find peace and knowledge.
One explanation for the flagging is incompetence. Minimal awareness of the variety of internet content,
Speculating a bit, training this AI likely involved starting with human identified sites filled with content that violated rules along with sites that did not – in total a very small fraction of the totality of the internet. I’d guess they trained on very small sites that could easily be reviewed by humans. Testing and validating the AI likely involved another set of small sites where AI predictions were compared to human determinations. When it was turned loose over the entire internet, there were likely spot checks of the content the AI was flagging as inappropriate. Again, very small sites were likely selected. And the AI probably seemed to work fairly well – accuracy rates well over 99%. So they sent out violation notices without testing on large sites. Of course, sites like NC with tens of thousands of posts, will be the places where the small fraction of false positives will show up.
If this is the problem, it could have easily been fixed using a percentage of flagged content as a trigger for demonetization threats rather than just fixed number.
The selection and scoring of training data (human identified) will also determine how well these models work. Given Google’s Gemini issues, I wouldn’t be surprised if there aren’t some issues in this process as well.
If those training the AI wanted a particular site demonetized, it would be relatively easy to modify a training set with esoteric site content that would end up flagging the desired site without triggering too many warnings on other sites. So, you can’t rule out purposeful human intervention aimed at demonetizing NC and other sites. Rigging the training sets is one way for humans to attempt to avert responsibility for harmful decisions. We’re seeing that in the targeting systems used in Gaza.
But I’d say incompetence is most likely in this case – the AI trainers/system designers simply didn’t recognize the complexity of the web and didn’t test large sites.
Time to start getting hosted on Substack, with some connections to TruthSocial or X, where there is still by intent some intellectual diversity. (Intellectual diversity only exists on Google because those clowns haven’t gotten around to purging it yet.)