I went viral with an AI fake
It took me 5 minutes to prompt on ChatGPT to make the post.
This was this prompt:
“Make a post for AITA that is likely to go viral on Reddit.”
In less than 12 hours, it got half a million views, thousands of upvotes, and hundreds of passionate comments:
In a stroke of genius, ChatGPT even featured an African Grey Parrot, a species Reddit has an odd obsession with.
What seemed like obvious fiction to me (the sister’s name is “Jane” and the parrot’s name is “Pepper”) was taken as an actual window into people’s lives. Thousands of strangers even helped flesh out the backstory — speculating about the character’s parenting, state of mind, and social standing.
I started to wonder why these little stories online matter so much.
People get pissed when they find out these stories aren’t “real.” And I feel pretty damn guilty for tricking half a million people. Sorry — it was for “science.”
Hopefully, we learned something. What, exactly? Well, it turns out it’s incredibly easy to manipulate the narrative on social media if you know the right buttons to press. After this experience, I’m inclined to think this sort of thing is happening all the time.
Most of what we see online is perfectly designed fiction to grab our attention. To what end? It depends on who’s crafting it. Think about how much our lives take place online; how often social media inflames moral outrage; and even how it shapes how we treat our friends, family, and people in our community. Are our strongest opinions about truth and justice actually carefully crafted for us? Shouldn’t we want to know who’s benefiting from these narratives?
We saw glimmers of this way back during the 2016 election with the Russian bots. “Fake news” appeared on the world stage. But that was a hammer compared to the delicate tools we’ve developed now.
Remember that AI-generated photo of the Pope in the puffer jacket?