Daily Reminder To Not Listen To Google’s AI Overview | Whatever
Whatever
DIGITAL CONTENT FROM ANALOG PEOPLE
Daily Reminder To Not Listen To Google’s AI Overview
Posted on August 19, 2026
Posted by Athena Scalzi
17 Comments
Athena<br>" data-image-caption="" data-large-file="https://i0.wp.com/whatever.scalzi.com/wp-content/uploads/2023/08/WHSAthenaS2.jpg?fit=225%2C337&ssl=1" class="alignright size-full wp-image-48640" src="https://i0.wp.com/whatever.scalzi.com/wp-content/uploads/2023/08/WHSAthenaS2.jpg?resize=225%2C337&ssl=1" alt="" width="225" height="337" srcset="https://i0.wp.com/whatever.scalzi.com/wp-content/uploads/2023/08/WHSAthenaS2.jpg?w=225&ssl=1 225w, https://i0.wp.com/whatever.scalzi.com/wp-content/uploads/2023/08/WHSAthenaS2.jpg?resize=200%2C300&ssl=1 200w" sizes="(max-width: 225px) 100vw, 225px" />Hey y’all, I’m here today with a PSA that I find myself repeating multiple times a day, every single day. And I’m going to do it again. AI is wrong a lot. And I mean a lot. Both about inconsequential things and extremely consequential things. But the fact it is so wrong so consistently about inconsequential things should make you think twice about listening to it regarding anything even remotely serious.
For today’s example of "AI is always wrong and I hate that it just lies so confidently right to everyone’s face," I’ll be talking about Connections.
I play Wordle and Connections almost every day, as do many of my friends. Today, for some reason, the Connections game board is comprised of red and white tiles. This is not something I’ve ever seen before, and none of my friends know why it looks so different, either.
So, I Googled, "why is the connections board red," and Google AI Overview told me "The Connections game board turns red when you make a mistake or get an answer wrong."
Guess what! THAT’S NOT FUCKING TRUE!
Please please please stop Googling things, looking at the AI Overview, and assuming it has given you the correct answer to your query. Don’t trust Google AI with anything, even with inconsequential things! Because it will, and DOES, lie to you. And it can’t be held accountable for giving you wrong information. It just does it and suffers no consequence. The consequence falls only on you.
If you know why the game board is red, please tell me, I am still very curious.
-AMS
Share:
Share on Bluesky (Opens in new window)<br>Bluesky
Share on Threads (Opens in new window)<br>Threads
Share on Mastodon (Opens in new window)<br>Mastodon
Share on Facebook (Opens in new window)<br>Facebook
Share on Reddit (Opens in new window)<br>Reddit
Share on LinkedIn (Opens in new window)<br>LinkedIn
Share on Pinterest (Opens in new window)<br>Pinterest
Share on Tumblr (Opens in new window)<br>Tumblr
Share on WhatsApp (Opens in new window)<br>WhatsApp
Email a link to a friend (Opens in new window)<br>Email
Like this:<br>Like Loading…
Category: Athena Scalzi
← The Big Idea: Tim Pratt
17 Comments on “Daily Reminder To Not Listen To Google’s AI Overview”
I find Google’s AI useful for little bits of research, but then I go to Wiki for more in-depth answers. At least Wiki makes efforts to keep articles accurate, like noting when a statement needs some kind of citation backup.
I think the Connections board is supposed to look like a classic red and white gingham picnic blanket as a callback to some of the clues.
At this point I just assume it is wrong all the time. I’ve turned it off by using udm14, because even seeing it accidentally spikes up my cortisol now. Straight up nonsense 🙄
Pretty sure that’s what someone who lost at Connections would say. At least, that’s what Google AI said.
The board is checkered to look like a picnic blanket. It’s part of a running theme through some other NYT puzzles today (Strands, Mini, Midi, Crossword).
Just a misdirection.
I will not say more as I don’t want to share spoilers ( or maybe I already have?)
I have also noticed the negative impact of AI on Google search. I have been using Google Search and Google Chrome for a long time but I just finished migrating to Firefox and using Duck Duck Go search and I’m only a week in but the browser seems faster and it’s surprising how much I appreciate a search returning links. I do use ChatGPT and Gemini for research but it seems when I searching for a website, I much prefer being the one who decides the resource instead of AI.
Oh, Athena.
Please read "On Bullshit" by Harry G. Frankfurt .
It describes the difference between lying and bullshit, and argues that bullshit is worse,
I have identified it as one of my biggest problems with DJT, and it is also the reason that I put about as much faith in Google’s AI as you do.
I always, always, always check the sources Google’s AI says it’s pulling the info from, go read those pages, and if I know it’s bullshit already, I look at the websites the regular search returns to make sure I’m not relying on bullshit answers from AI. Not just with Google, with any AI. And to round it...