Google’s 'Woke' AI Disaster Reveals a Dire Problem
Lies, damned lies, and… artificial intelligence
"The danger of training AI to be woke – in other words, lie – is deadly."
~ Elon Musk
In 2023, Elon Musk said the above quote about OpenAI's ChatGPT chatbot. It was around the time when ChatGPT became the fastest-growing consumer application in history.
Following ChatGPT’s success, AI stocks exploded.
But ChatGPT had a big problem. It was… well, let's just say it had a penchant for taking sides. And it wasn't just Musk who noticed.
For instance, the chatbot was all too eager to praise President Joe Biden and Vice President Kamala Harris, but it flat-out refused to say anything nice about former President Donald Trump. It also refused to talk about the advantages of fossil fuels.
The program also suggested it would rather detonate nuclear weapons, killing millions of people, than use a racist slur.
Fast forward to today, ChatGPT is coming up against some stiff competition from Google’s Gemini, a new AI platform even more biased and politicized.
The Woke Bot
Gemini is very similar to ChatGPT. Give it some prompts, and it’ll answer questions, generate images, maybe even help you cook up a song or a recipe. You get the idea.
Now, people began tinkering with its new image generation feature, and they soon noticed something quite peculiar: Gemini outright refused to produce or generate anything involving white people.
Request a picture of a happy white family, and instead of providing the image, Gemini would embark on a 400-word rant explaining why generating such a picture would be "exclusionary."
But ask for a picture of a happy black family, and you could get all the happy black families to your heart's content. Or Asian, Samoan, or anything else, as long as it’s not white.
And it became even more absurd when people got into history.
When asked to show pictures of the founding fathers, Gemini responded with images of black and Native American individuals signing what appeared to be a version of the U.S. Constitution.
When prompted for a picture of George Washington, Gemini generated an image of a black man in a white wig wearing an army uniform. Why the unexpected twist? Gemini simply said it aimed "to provide a more accurate and inclusive representation of the historical context of the period."
And it held no punches back trying to live up to this “noble” goal… you got everything from a Southeast Asian woman pope to ethnically “diverse” Vikings.
You don’t need to be a history buff to realize that none of that aligns with historical facts. Yet, Google’s Gemini boldly decided to fake history.
And if it couldn't fabricate the desired image, it simply ignored the request. For instance, when someone asked Gemini to show the events of Tiananmen Square, it provided a laundry list of excuses for why it couldn’t do it. This, of course, referred to infamous Tiananmen Square massacre of dissidents, orchestrated by the Chinese government in the late 1980s. Clearly, Google’s Gemini has something of a soft spot for China.
The internet, naturally, had a field day with this, poking fun at both Gemini's obvious biases and Google's fauxpology, claiming the feature "missed the mark."
But there was one bunch who weren't too pleased: Google's shareholders.
Google's parent company Alphabet took a hit of over $70 billion in market value in just one trading day on Monday as the controversy hit Wall Street. Ouch.
Gemini's image generation feature has since been paused, with plans to reinstate it in the next "few weeks.”
Peeling Back the Layers
Even the most sophisticated AI chatbot, deep down, is just a piece of software... designed and coded by real people.
So, it's no surprise that people quickly scoured Jack Krawczyk's X/Twitter account, the top executive at Google responsible for creating Gemini.
Suddenly, it all started to make sense... Jack’s wall was filled with statements urging people to feel guilty about their “f—king white privilege” and asserting that America is plagued by “egregious racism.”
I won't get into his other tweets (and there's quite a collection), you can look them up yourself. Suffice it to say, the guy has an axe to grind.
Now, when you see these images, sure, they're funny and ridiculous. But if you brush it off as merely a frivolous side issue, think again.
AI isn't just a passing fad or a phase we'll eventually shake off like a bad case of measles. Nope, it's a multi-billion-dollar industry led by giants like Google and Microsoft, and it's poised to revolutionize how we live, work, and do business for decades to come.
One of the earliest AIs was a program that learned how to play checkers back in 1952.
Since then, AIs have done it all – translating languages, mastering driving cars, becoming the world's best chess players, discovering new drugs, and even guiding unmanned probes to Mars.
Nvidia, the face of the AI revolution, now matches the combined value of all Chinese stocks in Hong Kong. Its worth has skyrocketed by $1 trillion in just the last year.
The AI market is projected to surge to nearly $2 trillion by the end of the decade, a 9-fold increase from around $200 billion in 2023.
Bottom line. The AI trend isn't slowing down anytime soon. And the technologies it brings will become deeply woven into our society's fabric.
This brings me back to the Elon Musk quote I posted earlier. Teaching AI to lie is potentially "deadly" precisely because of its immense potential, making it many times more dangerous than, say, a poorly designed aircraft.
As I write this, AI language learning programs are being introduced to schools. Google has already put a Google Chromebook with Google AI and Google search in the hands of almost every public-school student in America.
Just think about it. This is the knowledge base your children will use to learn about history, understand contemporary reality, and shape the future.
The tech giant is also rolling out AI across nearly all its products, from Android to the iOS Google app, Gmail, Docs, and beyond. Soon, it will extend to Search and Ads too.
Remember my earlier example with Tiananmen Square? If Gemini refuses to generate an image of what happened there, you can still find those images online because they exist in the real world. But what if this image-making tech merges with Google proper? Imagine searching for images of Tiananmen Square on Google and getting the same response: "We don't feel that these images accurately represent what happened at Tiananmen Square, so we will not be showing them to you.”
For someone who grew up in a pretty unfree place like the Soviet Union, this feels all too familiar. After all, it was the Soviets who mastered the art of image manipulation, trying to scrub certain figures from their history books. You could say they were ahead of their time (which might explain why Gemini seems to be taking cues from them).
But, of course, censorship didn't stop with images – it went far beyond. The Soviet apparatchiks took it upon themselves to meticulously (and unscrupulously) control the narrative they wanted the public to see. It was akin to this famous quote by George Orwell from his magnum opus, "1984":
“Every record has been destroyed or falsified, every book rewritten, every picture has been repainted, every statue and street building has been renamed, every date has been altered. And the process is continuing day by day and minute by minute. History has stopped. Nothing exists except an endless present in which the Party is always right.”
I'm not saying history will repeat, but it may just rhyme.
Until next time,
Lau Vegys
Makes me want to collect history books now to preserve for future reference in order to defeat new foolish narratives.
Just goes to show: every artificially 'intelligent' machine has a human asshole on the driver's seat.