As I mentioned yesterday in my Sunday Smiles essay, I think the hilarious failures of Google’s AI Gemini were one of the week’s most important stories.
That may seem odd since the first impression that many people get is Gemini’s odd preference for portraying Vikings as Blacks, Popes as minority females, and our Founding Fathers as minorities. What could be so important about that? Google will rework the algorithm and the problem will be fixed.
It likely already is. Sort of.
Gemini’s ridiculous–I mean that literally, as it inspired an enormous amount of ridicule–photo generation bias isn’t really important in and of itself. AI photos generated by Gemini are not good enough to be “deep fakes,” and everybody can see how stupid these images look.
So why is the story so important?
It showed us the man behind the curtain, and not in some abstract way. We got to see the world Google is trying to recreate in pictures worth more than a thousand words.
We see that Google’s world is a carefully constructed lie.
🚨 Gemini’s Problem: Google’s LACK of diversity
There is a broad assumption that Google Gemini’s clearly biased results are an artifact of Artificial Intelligence gone wrong, and that the issues are combinations of bad parameterization, flaws with the algorithms, and other…
— The Yootopian (@TheYootopian) February 24, 2024
Chances are that you already knew that Google carefully curates the information that you are given and that it biases its results for the benefit of the Left. I wrote last Friday about how Google News actually rewrites your requests in a manner that generates Leftist results. You ask one question and get results to another Google decided you should have asked.
Google Gemini revealed to all the world just how biased its search engine is.
The Yootopian, an account on Twitter that I follow (and you should too) wrote an essay that I think was spot on about how Google could blunder so badly in a major product launch.
As Yootopian pointed out, there is precisely zero possibility that the people at Google just happened to miss a bug like this. None. Zip. Nada.
Consider: AI is the next big wave of money-making products to come out of big tech, and all the companies are pouring billions into it. Gemini isn’t even new; it is an iteration of Bard, their prior AI product, and no doubt was extensively tested.
So how did this happen? Nobody at Google saw the bias, even though the moment it debuts this PR disaster happened. And there is more to come. Gemini might even get Google sued for defamation. It basically accused a prominent Twitter account owner of sexual assault, and referred people to the sexual assault hotline.
Yootopian provides what I believe is the correct answer: the Google employees who produced this disaster DID NOT SEE the problem. They are so ensconced in their comfortable bubble that they thought Gemini’s answers were perfectly fine. It is exactly how they see the world, so when it got out into the wild and people who don’t see the world as they do got ahold of it hilarity ensued.
A Google employee sees an answer calling Robert F. Kennedy Jr. hateful and just nods. He is hateful, racist, or otherwise discriminatory, so he should be silenced. It literally never occurred to Googlers that others might disagree, so when they tested the product out the answers seemed obvious to them.
As Yootopian put it, succinctly and well, Gemini isn’t failing to perform as expected. It is doing exactly what Google wanted it to do: present the world in this bizarrely distorted way, because this is how Googlers see the world.
There is a broad assumption that Google Gemini’s clearly biased results are an artifact of Artificial Intelligence gone wrong, and that the issues are combinations of bad parameterization, flaws with the algorithms, and other technology root causes. But the fact is: Gemini’s problems are NOT technology related. Gemini is performing completely as expected. This is what it was meant to do. Keep in mind that Gemini has been in development for nearly a year, and there is no doubt that it has been heavily tested. Google has seen these results for months (at least) and believed they were completely normal. As mentioned: to employees at Google, it was performing AS EXPECTED.
In my judgment, this analysis is exactly right. The problem isn’t with Gemini. It is with Google itself, and the fact that their “diverse and inclusive” employee cohort is so ideologically uniform that they see the world in exactly the same way, with exactly the same blinders, and are hence utterly incapable of seeing things that are obvious to everybody else.
Any given video of Tucker Carlson will get 30-40x as many viewers as a Rachel Maddow segment, and in the Googleverse this is a disaster. He must be silenced, so Gemini does. Melania Trump, who is one of the least offensive people in the world, whatever you think of her husband, cannot have a voice. Jill Biden? She’s great! Dr. Jill, who should be the Surgeon General!
It’s inarguable that the tech industry is dominated by liberal ideology. Political donation data from the FEC provides clear evidence (from the 2018 midterms): 96% of Google employee donations went to Democrats (and most other large tech companies were well above 90%, too)*. Furthermore, ex-Googler James Danmore’s “Google memo” documented how this one-sided view was actively fostered by leadership – an observation for which he was fired (which helped make his case, IMO)*.On top of this, Diversity, Equity, and Inclusion (DEI) initiatives that are meant to increase “diversity” often end up making companies LESS diverse because they promote physical characteristics rather than diversity of THOUGHT. Strong DEI proponents are nearly always left-leaning politically, and therefore corporate programs are often geared to support and attract people with similar ideology. Conversely, people who question the effectiveness of DEI initiatives are usually right-leaning politically, so they resist working at such organizations or just hide their opinions knowing that dissent is not tolerated (ask conservatives at a tech company if they are comfortable sharing their views – the response will be a near-unanimous “no”). Danmore found out the hard way.The result of all of this: an echo chamber full of politically left-leaning head nodders.
What makes Gemini’s ridiculous rollout is just how ridiculous it is, in the sense that it has generated so much ridicule.
Ridicule is far more damaging and revealing than mere analysis or criticism. I can write 5000 words, have them read universally, and have less impact than this photo:
I wish that weren’t so, since I have a big ego that needs to be stroked constantly, but it is. The saying that “a picture is worth a thousand words” grossly understates the impact of pictures, and Google’s AI-generated pictures are devastating to its credibility.
Everybody can see that Google, to be blunt, lies to us about what the world looks like. To every defender of Google, we can retort with a stack of absurd photos.
Defend this. It ain’t no algorithm issue. Google wants to give us its version of reality, not reality itself.
There is nobody with an IQ above room temperature who doesn’t believe that this was just a mixup with the algorithm. Everybody knows, at least in their hearts if not their heads, that Gemini got here because of DEI.
Google will hardly be destroyed by this debacle, but it will hurt more than a lot of people think.
We have seen the man behind the curtain, and the Great and Terrible Oz is a guy who thinks the Pope is a black woman.
