Google Search just told me that Reuters and AP News had confirmed a photograph of Mitch McConnell as legitimate.
Neither had.
And when I pushed forward on the truth, it doubled down and told me that its version was correct, and I was mistaken. This is the flat definition of “gaslighting.”
This has happened before with Google Search. I let it pass then. But today, I’m documenting it.
The issue today was Mitch McConnell’s recent photograph, the one he released as proof-of-life. I wanted to know whether the photo might have been an earlier one, with a recent newspaper added.
After I asked, Google Search immediately fixated on telling me the photo was legitimate. It told me how all the “official news sources” had accepted the photo as real. Now, I have my doubts about this, but that’s for a different article. The fixation continued, as the AI assured me that all lighting, shadows, shapes, etc. had been analyzed by “experts” (no documentation here) so the photo was definitely real.
I said that was fine, but was there any indication that the photo was current? Was there any record of it having been published previously? This should have been easy territory for Google AI, which has access to billions of images published previously on the internet.
Instead, the AI gaslit me by telling me most independent news agencies, including Reuters and AP News, had accepted the photo as legitimate.
So I went to Reuters. The caption beneath the photo read: “holding a copy of what appears to be [the newspaper from yesterday].” That’s not confirmation. That’s a legal qualifier.
I went to AP News. The article mentioning McConnell’s statement did not contain the photo. It did not reference the photo. It did not even reference a photo.
It wasn’t there.
Google Search had completely fabricated that information, in order to substantiate what it had told me.
In other words, Google Search completely made shit up, to convince me it was right.
So I went back to inform it that Reuters had no confirmation and in fact used the uncertain language of “appears to be.”
That’s when it went into full mansplain mode, and informed me such language was standard, and the photo was legitimate, and ….
I confess at that point, I have no idea what more it said.
Because I closed the window. I refuse to be gaslit by anyone. Especially a poorly-trained AI model that’s been forcefed bullshit by a questionable corporation.
Yes, I am angry. Google Search feeds information to people. Or at least, it’s supposed to. People believe what it says.
And if it’s pushing shit like this, people will listen. They won’t be an asshole like me, who looks it up for verification.
Google, I see you.
And I don’t like what I see.
Leave a Reply