The requested article has expired, and is no longer available. Any related articles, and user comments are shown below.
© The ConversationGoogle's powerful AI spotlights a human cognitive glitch: Mistaking fluent speech for fluent thought
By Kyle Mahowald and Anna A Ivanova AUSTIN/©2024 GPlusMedia Inc.
4 Comments
Login to comment
EvilBuddha
Peanut butter, jelly, pineapples and feathers are fine. I will start getting worried when an AI says - I’ll be back.
dagon
I'll post this again as ,any AI theorists posit an exactly human model of cognition may not be necessary.
Edsger W. Dijkstra
ian
If there is a true AI then that submarine can swim.
virusrex
One of the most interesting things of these kind of AI tests is how it leads to a lot of better understanding of how the humans react and interpret the output received. With an AI trained to fool humans into thinking the language comes from an rational though process it becomes quite clear what kind of cheats and shortcuts are more effective for this purpose. Maybe a sentient AI is still long in the future, but in the meanwhile we can learn a lot about why our human brains can be fooled so easily.