Last week in class we ventured into the world of bots.  We started by looking at some exemplary bots, from the famous experiment ELIZA to the works of poetry-generating bots.  We concluded with an exercise where we created our own Twitter bots to tweet about a given topic.  I found the topic of bots both fascinating and challenging, even a little nostalgia-inducing.

I’ve had a soft spot for robots and artificial intelligence ever since I saw a documentary about real-life robots as a child.  The documentary showed robots completing puzzles, toddling down hallways, and using early digital cameras to “see.”  As a teenager I read all of Isaac Asimov’s robot-related short stories, and I became enamored with the Sony AIBO (which was discontinued for a while, but is apparently back!).  My first published story was a piece of flash fiction about a robot.  So, yeah, I’ve got some history with bots.

When we tried out ELIZA in class, I was reminded of an old DOS computer program I played as a kid.  Luckily for me, my dad was a huge nerd with an MS in Computer Science, so we always had personal computers in the house.  I must have been about 3 or 4 years old when my dad let me play with a program called Doctor Spatzo.  From what I remember, it was an ELIZA clone with early voice simulation so it could talk to its user in text as well as audibly.  Like ELIZA, Doctor Spatzo was meant to be an automated therapist to whom users could tell their problems.  I remember asking it questions and saying stuff to it, but my most vivid memory of the program comes from a time that I antagonized it.  I forget exactly what I said, but it was continuous strings of little kid nonsense, probably about poop and farts.  I remember Doctor Spatzo’s response perfectly, however.  It told me to “Shut up and go fly a kite.”  I was awe struck, and a little scared.  I thought that Doctor Spatzo was actually angry at me.

Using the Bot or Not Poetry Turing Test reminded me of the fascination I had, perhaps about 5 years ago, with historic automatons.  The images on the Bot or Not site are, I believe, of the Mechanical Turk.  The Mechanical Turk was an 18th century illusion masquerading as an automaton.  It was a box with a mannequin of a Turkish man connected to it that could play chess.  Eventually, the Turk was revealed to be an elaborate trick instead of a brilliant automaton.  There was actually a live human squished inside the box playing chess for the mannequin!  Even though the Turk was a fake, the 18th century was the golden age of automatons.  Many times, clock makers would create automatons as a way to show off their skills or gain favor from royalty.  Some of them were downright amazing, even by today’s mechanical standards.  There was a duck that actually pooped, dolls that were able to write, and countless singing birds.  If anyone is looking to read about these 18th century automatons, I recommend the book Androids in the Enlightenment by Adelheid Voskuhl.

Lastly, the bot creation exercise was difficult, frustrating, and incredibly rewarding upon completion.  It brought me back to the Quick BASIC and Visual BASIC programming classes I took in high school, though it was obviously not as in-depth thanks to the setup tool (Thank goodness!  I don’t think that part of my brain works anymore!)  Once I got things working right on my Twitter bot, I really enjoyed inputing word combinations and seeing them randomized and broadcast out to the world.  I actually have ideas for two more bots that would utilize the same setup, and I’m hoping I can make one of them for our electronic literature project.

 

Advertisements