I’m a casual gamer. That means that I like video games, and play them, but I’ve never been one of those Call of Duty, headset on, teabagging my dead enemies on the ground, kinda guy. Regardless, I think video games are an amazing art form that I’ve had the pleasure of seeing evolve over the course of my life.
I’ve always been a fan of story-telling, and ultimately making people care about something that I (or someone) made up. While not all games are this way, the video game industry has become an excellent medium to tell a story. I’ve always enjoyed narrative-based games, and in fact, my final paper in college was refuting Roger Ebert’s claim that “video games are not art”.
But I digress, I would love to go into the emotional adventures video games can take you on, but that’s for another blog post. I brought up games because my completion of a great one recently has inspired me to write about the Turing Test.
Detroit: Become Human is a 2018 video game for PlayStation 4 that is set in the near future and tells the story of an android population becoming conscious beings, and rising up to gain human rights and free themselves from the slavery they are under.
In this game there is an android character named Chloe, and in the game, she was the first android to successfully pass the Turing Test, starting the beginning of the production of androids and the entire story.
Great…..so what? Who is Turing and what is his test?
Alan Turing was this guy.
A British computer scientist and mathematician who was integral in the Allies’ cracking intercepted enemy messages that led to several major Nazi defeats during WWII.
Many would refer to Turing as the father of modern computer science, and his contributions to the field are massive and in excess of this blog. The one I want to cover, is the Turing Test.
The Turing Test was developed by Alan Turing in 1950 and its purpose was the test a machine’s ability to respond and behave in a manner that would be equivalent to, or indistinguishable from the behavior of a human being.
A human evaluator would be made aware that of the two entities he or she would be communicating with, one was another human, and the other was a machine designed to sound like a human with its responses. All three participants would be put in separate rooms, and the entire conversation would be limited to text. This allowed for the computer’s inability to sound like a human to interfere with the experiment. If the evaluator is unable to confidently say which was a human and which was a machine, the machine is said to have passed the Turing test.
The Turing test primarily is a scale of how immersed a piece of software can be in our lives. If it’s possible to have a digital secretary that only exists as a program made up of millions of lines of code, people will use it, but only if its language sounds like that of a human. It’s something all humans need, other humans.
Ultimately, machines being able to mimic human behavior is a major hurdle for the industry to conquer.
Human idiosyncrasies are unusual and hard to define, and as a result, are something computers have a hard time understanding; yet their use is more or less vital in the passing of a Turing test.
Products like Google Duplex AI and Bowtie.ai try to make use of computers playing the imitation game and disguising as humans.
Google Duplex is an AI software that can make appointments over the phone for you. In a presentation, the AI booked an haircut appointment with a human that was none the wiser on the other end of the call. At a certain point during the call, the AI responds with “mm-hmm”, instead of “yes”. Those little things are how a Turing test is passed; reflecting human slang and diction.
One of the biggest challenges facing autonomous cars are things like the pedestrian on the street waiving right-of-way and waving you through. Sometimes even brief eye-contact can communicate a lot with a human that a computer would have a ton of trouble doing.
It’s hard to not fantasize about the future with all the media we have flying around, not to mention the rapid speed at which technology grows. The original TV show “Lost in Space” was set in the year 1997, so it’s not uncommon for us to overshoot when thinking about future technologies.
Regardless, the more programs and machines can sound like humans, the more integrated they will become with people’s lives, and the more people will want to use and buy those programs. So it seems inevitable that AI passing the Turing test will become more common.
And who know, maybe someday we’ll have a machine that can pass a visual Turing test, gain consciousness, and start a revolution.