A Chatbot Named Eugene Tries to Cheat on the Turing Test
This week, the press was all over the story that a Russian-developed chatbot posing as a 13-year-old Ukrainian boy named “Eugene Goostman” was the first instance of an “artificial intelligence” to pass the Turing Test. The Turing Test, if you’re not familiar with it, is a test designed to measure the “humanity” of an artificial intelligence through text-based conversations. The benchmark: It has to fool humans into believing that a person is doing the typing.
The tenor of the first news reactions: In less time than you can say Bletchley Park, we went from “Wow, isn’t technology amazing?” to “Oh my god, we’re all going to become a race of slaves to our new robot overlords!”
But don’t bow down to those overlords just yet. I have read some of Eugene’s answers to the judges and have spent some time chatting with him myself. I am not worried.
Yes, it will be interesting to see how the real field of artificial intelligence develops. But I can say with confidence that we are all safe for now.
As the mother of a 14-year-old boy with whom I attempt to converse on a daily basis, I can tell you that Eugene is way too chatty to be mistaken for your average male teenager, whose conversational output is usually measured in grunts. Plus there wasn’t a single “lulz” or “POS” (“parents over shoulder”) in the whole exchange. So right away I knew Eugene wasn’t real.
In order to explain what really happened in the 2014 Turing Test, we thought it would be fun to take a peek at Eugene’s home life, to see what happened when he told his parents about his big achievement.
Scene: The home of Mr. and Mrs. Goostman, computers
“Mama! Papa! I’m home!”
“Evgeny! Sit, eat, I made fresh data for you. So, tell me, how was your day?”
“You’ll never guess what happened! All of Bletchley Park Middle School had to take a test, and I’m the only one who passed!”