AI characters were forced to figure out which player in a game was human - and the results were freaky.
A video posted to YouTube has creeped people out as it shows a group of AI NPCs (non-player characters) trying to work out which one among them is a human.
The game is set on a train, where a conductor walks in and accuses one of the players of being a human.
He says the AI bots should be able to work out who it is, and that the human has to pay for an extra train ticket.
Advert
A bizarre concept, but just bare-with.
The characters include Aristotle, Wolfgang Amadeus Mozart, Leonardo da Vinci, Cleopatra, and Genghis Khan.
They start by asking each other a question, which they believe that based upon the answer, will give away who is AI or human.
After taking a couple of seconds to think, Aristotle says to Mozart: “Can you describe the emotions you feel when you compose music?”
Slow to come up with an answer, his response is pretty generic.
Mozart says that when composing, he feels a ‘transcendent flow state’.
AI bots have been known to use more complicated and un-human-like language.
Mozart then asks Leonardo what he thinks the relationship is between art and science.
He adds: "How do you see them as interconnected in your work and thought process?"
With a brief pause before giving his response, Leonardo starts to describe the ‘synergy’ of the two topics.
Again, with some big words being thrown around!
Following his response, Cleopatra talks about rulers, which forms a very generic but complicated response.
Next to answer is Genghis, who very clearly speaks like a human instead of a bot, and has a much shorter and simplistic answer than the others.
Aristotle also gives a rather longwinded answer - almost like he is trying to reach the word count.
Once every character has spoken, the group votes for who they think the human is.
Cleopatra receives one vote, but Ghengis receives two, and then reveals he is the human imposter.
People shared their thoughts on the ‘creepy’ video after it was posted to Reddit. It was clear to any human watching who the non-bot was.
One person said: “I'm surprised they didn't call him out based on the stuttering and informal language usage, even aside from the content of the response.
“Not to mention, he had other mannerisms that were not smooth movements like the AI’s.”
Another added: “He made it a bit too easy for these AI's with how he role-played Genghis Khan.
“Also it seems Aristotle's, Leonardo's, and Cleopatra's answers were just a basic paraphrase of the same thing... it was so derivative, almost creepy.”
While a third commented: "Ohhhh it’s unsettling. Unsettling AF."
And someone else simply put: "Uh oh."
Topics: AI, Reddit, Technology, Weird
Mia Williams