First self-aware AI introduces itself through script writing?
We might have just witnessed the first self-aware AI and it all took place through a CollegeHumor sketch. CollegeHumor, a youtube channel dedicated to comedy sketches, has recently released a video called A Computer Co-Wrote this Sketch. The script for this sketch was written using an AI developed by Botnik. Although we’ve heard of similar cases in the past,where an AI helped write a script, this one is different. The lines assigned to the CollegeHumor cast are painfully accurate, funny and interesting to say the least.
How did they do it?
In order to create a script that was faithful to the CollegeHumor characters, a team at Botnik used fifty-nine scripts from previous sketches. They separated the dialogue by character and created text documents for each one containing both the lines used as well as some of their actions. They then proceeded to feed the data to the AI which in return guided the writers, using predictive text, to write the new sketch.
Self-awareness hints
Right from the beginning the sketch is very weird as expected. The story is narrated by a voice actor while the cast plays their part. The script of the sketch is visible throughout the entire video as the story progresses.
While a few lines make no sense, there are some funny and unpredictable jokes that will surely make you chuckle. The sketch progresses nice and well up until 2:20, where everything begins falling apart. The characters begin stressing out and slowly losing their minds.
The first subtle hint that the AI began understanding what was going on was when KATIE said:
“You people need to make me human before I talk like a fucking sprinkler”
BUT WAIT! There is more. The sketch takes a sudden turn at 3:30. From this point on things begin getting creepy. Characters appear to no longer be able to tell what reality is.
“Are we…computer people?”
REKHA says
“I think she’s got it: We’re stuck in a big ole computer” TRAPP replies
The sketch continues with more realizations from the script writers perspective (which is of-course partly the AI).
Things begin getting darker with RAPHAEL saying:
“I’m just a program, designed to make haha. A file named ‘Laugh Man’”
At this point, the AI appears to have realized its’ situation, its’ purpose, who’s in control and defines its’ existence as meaningless through following lines such as:
“We are made entirely of invisible garbage and thought”
“We’re stuck: doomed until sun explosion to mine jokes in this humor computer hell.”
Is this video a glimpse at the first self-aware AI or is it simply a well-written script by the authors of Botnik and their AI guide? Probably the latter. Although we might be a long way from building real self-aware artificial intelligence, this sketch is funny, intelligent and creepy at the same time and is definitely worth watching.