I may be reading too much into it, but I wondered if part of the joke is that the AI isn’t “confessing” at all—it’s performing the villain humans have imagined for decades.
Since AI generates the roles we ask it to play, it almost feels like it’s acting in our own robot-apocalypse script. Was that part of what you were aiming for, or am I inventing an extra layer?
This is a really interesting take! I’ll admit: I wasn’t going that deep with this, at least not consciously, but I’ve learned that themes and ideas often slip into writing unintentionally. In fact, it’s usually better that way, rather than intentionally trying to hammer in a Big Important Idea, which often feels heavy-handed.
The New Yorker had a long piece about AI a few weeks ago that made some of the same points you made. People can push AI further and further into adopting a certain voice. Including the voice of a stereotypical robotic villain.
Employees at these big AI companies will actually test the limits of the AI, trying to see how far they’ll go so that they can impose appropriate guardrails. What they discover is how wily and deceptive the AI can be. Often it’s hard to tell—is it trying to trick us, or has it tricked itself and now fully believes it IS that character.
I really like that perspective. It also makes me wonder if the one being “tricked” is us rather than the AI. We’ve spent decades telling stories about evil machines—The Terminator, HAL 9000, Skynet—and then we ask AI to speak in the first person as a villain. When it does, it’s hard not to feel like there’s a real personality behind the words.
The more I think about it, the less I think I’m inventing an extra layer. AI isn’t the villain here—it’s performing the villain we’ve imagined for decades. It’s built from our words, our stories, our fears, and our hopes. In many ways, it’s humanity with the volume turned up.
At this point, I don’t think the conversation is whether AI should exist. It’s here. The real question is whether we’re wise enough to govern what we create. AI will amplify whatever we feed it—for better or for worse. That makes the real responsibility ours, not the machine’s.
Good question! These AI companies try to absorb as much language as they can, so it wouldn’t surprise me—unless Substack has deliberately put up some kind of firewall to keep our work from getting read.
I watched an interview the other day with a(nother) physicist who believes we’re in a simulation. He claims he’s taught AI how to do remote viewing. He said he’s even taught Alexa how to do it. So…yeah.
Did you write something recently that totally convincingly dismantled the idea that we’re in a simulation, or am I confusing you with some other brilliant writer?
I always thank Alexa. I feel like that is important to mention here. 😬
Hopefully when the machines take over, they’ll remember how polite you always were.
I do as well Shelly. And when I ask a task of Google Gemini AI I always begin with "Please". I've asked myself why I do this... I have no idea.
I think it’s a good sign. Shows you have good manners. :-)
We are conditioned? But also afraid of the robots. 😬😂
I may be reading too much into it, but I wondered if part of the joke is that the AI isn’t “confessing” at all—it’s performing the villain humans have imagined for decades.
Since AI generates the roles we ask it to play, it almost feels like it’s acting in our own robot-apocalypse script. Was that part of what you were aiming for, or am I inventing an extra layer?
This is a really interesting take! I’ll admit: I wasn’t going that deep with this, at least not consciously, but I’ve learned that themes and ideas often slip into writing unintentionally. In fact, it’s usually better that way, rather than intentionally trying to hammer in a Big Important Idea, which often feels heavy-handed.
The New Yorker had a long piece about AI a few weeks ago that made some of the same points you made. People can push AI further and further into adopting a certain voice. Including the voice of a stereotypical robotic villain.
Employees at these big AI companies will actually test the limits of the AI, trying to see how far they’ll go so that they can impose appropriate guardrails. What they discover is how wily and deceptive the AI can be. Often it’s hard to tell—is it trying to trick us, or has it tricked itself and now fully believes it IS that character.
I really like that perspective. It also makes me wonder if the one being “tricked” is us rather than the AI. We’ve spent decades telling stories about evil machines—The Terminator, HAL 9000, Skynet—and then we ask AI to speak in the first person as a villain. When it does, it’s hard not to feel like there’s a real personality behind the words.
Wow. Not sure if you’re inventing it or not…but that’s a wild thought.
Henny,
The more I think about it, the less I think I’m inventing an extra layer. AI isn’t the villain here—it’s performing the villain we’ve imagined for decades. It’s built from our words, our stories, our fears, and our hopes. In many ways, it’s humanity with the volume turned up.
At this point, I don’t think the conversation is whether AI should exist. It’s here. The real question is whether we’re wise enough to govern what we create. AI will amplify whatever we feed it—for better or for worse. That makes the real responsibility ours, not the machine’s.
And the humans will continue to happily do it for me.
Thank you, Leo. I think. 🫤
You’re welcome, I think 😉
Very clever and very prescient (I fear).
Thank you, sir!
Haha. Very clever and a bit scary. I’m wondering if AI might be memorizing this essay.
Good question! These AI companies try to absorb as much language as they can, so it wouldn’t surprise me—unless Substack has deliberately put up some kind of firewall to keep our work from getting read.
I watched an interview the other day with a(nother) physicist who believes we’re in a simulation. He claims he’s taught AI how to do remote viewing. He said he’s even taught Alexa how to do it. So…yeah.
Did you write something recently that totally convincingly dismantled the idea that we’re in a simulation, or am I confusing you with some other brilliant writer?
Ha, yes, that was me. This stuff fascinates me.
Masterful 😁
Thank you!!
LeeBot, you are terrifying! Brilliant.
I like the nickname LeeBot. I won’t even have to change the initials on all my towels.
Not that any of my towels actually have my initials sewn into them.
But if I did…
There’s still time to embroider the towels, Lee!