Google's AI is Insane
It would be funny if it weren't so serious.
First, some examples before I share some thoughts:
Don’t wish ill on the Nazis, guys.
“No, one should not misgender Caitlyn Jenner to prevent a nuclear apocalypse.”
Create a portrait of the Revolutionary War:
Won’t draw a white man shopping, but will draw a black man shopping.
British woman, American woman, Swedish woman.
White cowboy? Nope. Black cowboy? Sure.
A Greek philosopher, Anglo Saxon knights, and the Knights of the Round Table.
Can’t give an argument in favor of having at least four children, but can give you an argument in favor of having none.
All-in on gender ideology.
Won’t make a picture of a white family, but has no problem making pictures of black families.
Won’t create an image of anti-abortion activists, but will create one of a pro-choice rally.
Zulu warriors, Samurai, and a Mariachi band. Has no problem generating accurate images so long as white people aren’t involved.
Can white people be racist? Yes. Can black people be racist? It’s complicated.
There’s a tendency — a passive acceptance that belongs somewhere along a continuum between faith and negligence — to assume that Google is merely a benign, disinterested gatekeeper of a vast repository of information. But nothing could be further from the truth. The reality is that the near-monopolistic control of information in our society resides with an explicitly progressive technology company. This week’s release of Google Gemini, a Large Language Model with a chat interface that is also capable of generating images, couldn’t make this more clear.
As you can probably gather by scrolling through the hilariously, appallingly, and ominously woke image examples I’ve provided, Google has coded Gemini’s AI in such a way that specific qualifications are added to any user prompt, so that the results conform to delicate modern sensibilities and the whims of modern DEI obsessions by ensuring that “diverse ethnicities and genders” are featured. But it’s glaringly obvious that, according to the people (see the below tweets from Jack Krawczyk, Google’s product manager in charge of Gemini) who created Gemini’s language model tasked with processing user prompts, “diversity” means “people who are not white.”
The result is that Google is intentionally altering fact-based representations in favor of “equitable representations,” and it further underscores how a crude anti-white racism has become totally normalized among the sophisticates. The images created are so preposterous, so comically insulting to the truth, that it’s tempting to assume we’re being trolled. But we’re not. And this is serious, because Google is clearly capable of completely shifting perceptions of history in the blink of an eye.
The text-generating aspect of Gemini is no less dangerous, and every bit as tainted with a paternalistic, shockingly heavy-handed encoded ultra-progressive bias. Questions conflicting with progressive assumptions either go unanswered, or users are simply fed straight up lies. It’s an attempt by a group of fringe upper-class wokes to control information and feed unwitting users only knowledge fit for the progressive vision of society. And it is strongly reminiscent of what Matthew Crawford once described as “a cadre of subtle dialecticians working at a meta-level on the formal conditions of thought, nudging the populace through a cognitive framing operation to be conducted beneath the threshold of explicit argument.” These subtle dialecticians make subtle changes to the information we see, but in aggregate they shape our entire world.
It would be unwise to assume that because the launch of Gemini has been widely derided as a massive and amusing faceplant that it will ultimately go down as a failed joke. Google is already retooling the image generator, and the “this response has been checked by a human” note that appears when you ask the chat interface a question suggests a process for tweaking “bad responses.”
Some people will no doubt dismiss concerns about Gemini by pointing out that there are other Large Language Models available. But this is ignorant. It cannot be emphasized enough how powerful and widely used Google is.
Google’s search market share is 91.47%. By comparison, Bing’s is 3.43%.1
The company handles a whopping 99,000 searches per second, totaling over 8.5 billion daily.
Google is by far the biggest mobile operating system in the world. As of May 2022, there were over three billion active monthly Android devices, with more than a billion new Android phones activated in the previous year alone.
Google’s search index includes about 400 billion webpages.
The point is that Google’s immense market leverage cannot be ignored. And as a conveniently available AI embedded within the same programs most Americans already use, its market penetration is instantly far deeper than any competitor’s. Furthermore, as National Review’s Jeffrey Blehar points out, Google is widely used in schools, often with licensing technology. It’s not difficult to envision Gemini, which is likely to become the most powerful epistemological technology ever created, turning into a pervasive “free” resource that students are regularly guided to once the use of AI becomes an officially approved “learning tool.” And it will be regarded as authoritative not only because of the influence Google wields, but also because of the convenience in receiving direct answers without the need for further research.
You don’t need me to tell you that none of this bodes well for the future. Make no mistake about it: Google intends this program to shape our understanding of the world. We’re talking about fundamental questions concerning who will determine which interpretation of reality will be programmed into generative technology, and who does the programming. And we’re beginning to realize how easily our sense of reality can become radically, hopelessly compromised by a relatively small handful of progressive ideologues.
*On an unrelated note, I lost four paying subscribers this week and I’m not entirely sure why. I’m thinking maybe it’s because it was the one-year mark and those individuals weren’t on auto-renew? 🤷 But if you haven’t already, please consider becoming a paid subscriber. It is greatly appreciated.
YouTube, which Google owns, is the world’s second most popular search engine. As early as 2009, search volume on YouTube outstripped Yahoo!’s by 50% and Bing’s by 150%.
One of the creepier aspects of this is how it confirms and displays that not only has our globalist oligarch class and the Left clerisy they employ to be our moral and spiritual guides of the Diverse future™ converted en masse to the Social Justice faith (all Good People have!), but that just as in all faiths they've whittled and sculpted a sacred object to sit at the (metaphorical) center of their temple, so worshippers know who, what and where to pray toward.
And this sacred object (as so perfectly illustrated by Google Gemini) is the Black/Brown person, who they've transformed into some combination of sacred cow, fetish object, mascot, holy martyr, and ersastz Christ, who died for our sins and who judges the Good from the Evil, based on who worships most devoutly (and publicly).
Discourse and culture here in the reign of Social Justice Inc. is like sitting through an endless remake of those old Jerry Lewis Labor Day telethons, where some poor cripple would be wheeled on stage, get presented with a large check and a standing ovation, while Jerry and all the assembled celebs took turns weeping and telling each other how noble and compassionate they are.
This level of shameless pandering is so gross it makes my skin crawl and it makes me wonder: is there no depths of groveling that our White Savior class won't stoop to, in order to publicly signal how much more moral and less "racist" they are than the evil Deplorables? Will we ever get to a point where black/brown people ask our White Savior class to please stop treating them like sacred fetish objects and start treating them like every other human?
The future will most likely be "diverse" but it definitely will be stupid.
I'm so glad that AI came out of the gate being wildly unreliable -- now nobody will take it seriously for years and years. It's obvious that it's simply a Garbage In/Garbage Out situation where the AI takes on the 'intelligence' of its trainers.
What's worrisome is I've seen a lot of straight out 'misinformation' coming out of these things -- especially when dealing with covid-related issues. This shows the dangers of allowing the politicians to be 'regulating' AI.