Stanford University researchers paid 1,052 people $60 to read the first two lines ofThe Great Gatsbyto an app.
Much of the process was automated.
The researchers contracted Bovitz, a market research firm, to gather participants.
The interface reserachers used to make genrative AI agents.© Stanford University image.
as a way to calibrate the audio.
The two-hour interviews, on average, produced transcripts that were 6,491 words in length.
The researchers published the interview script and questions the AI asked.
This part of the process was as close to controlled as possible.
It then ran participants and the LLMs through five economic games to see how theyd compare.
They hit 80% on the BFI.
The numbers plummeted when the agents started playing economic games, however.
The researchers offered the real-life participants cash prizes to play games like thePrisoners DilemmaandThe Dictators Game.
In the Dictators Game, the participants have to choose how to allocate resources to other participants.
The real-life subjects earned money over the original $60 for playing these.
Faced with these economic games, the AI clones of the humans didnt replicate their real-world counterparts as well.
On average, the generative agents achieved a normalized correlation of 0.66, or about 60%.
Given time and energy, they can probably bring the two closer together.
This is worrying to me.
What happens when those machines have a script?
What happens when they have access to purpose-built personalities based on social media activity and other publicly available information?
News from the future, delivered to your present.
