You almost wish the upcoming biopic “The Imitation Game” starring Benedict Cumberbatch as Alan Turing would have been out right now to present some irony to the Turing Test being mentioned recently in the media. If you pay attention to science news recently, you probably heard about a new artificial intelligence program that supposedly passed the legendary Turing Test of AI supposedly becomes self-aware. It’s something Turing himself predicted 64 years ago would happen by the year 2000, and he’s perhaps only 14 years off. Then again, the Turing Test may have become a little too easy to pass by default.
The artificial intelligence program in question was nicknamed “Eugene Goostman” and done in the guise of a 13-year-old Ukrainian boy. Used in a test at the University of Reading in the U.K., it supposedly fooled about 33% of the people who used the program, which is just enough to pass Turing’s test for the first time. It was part of the Turing Test Prize that’s been held before in seeing how far artificial intelligence programs can go in fooling the person conversing with it.
But some in the media have questioned whether just because an AI program passes the Turing Test means it automatically equates the program with being self-aware. That’s a debate most of us thought would be coming years down the line and not so soon. It comes at a time when online artificial intelligence programs have already come close to passing the test.
One of those in the last couple of years was an online bot called Cleverbot that so many have claimed to be fooled by. Not long ago, I took on a chat with Cleverbot to see for myself whether it would fool me into thinking it I was conversing with a human brain. After conversing for half an hour, I realized that it was taking answers already used from other conversations and ended up giving too many non-sequiturs to make it truly comprehensible. Nevertheless, it still had a sophistication that no other artificial intelligence program ever had before.
Even a filmmaker recently decided to use Cleverbot last year create a screenplay for a short film. While the screenplay turned nonsensical as nothing more than a hilarious experiment, the creative imagination the bot provided was just the beginning of perhaps these programs providing create ideas in the future.
Now that we’re back to debating in movies whether letting artificial intelligence think on its own is a good idea, should we be thinking about whether AI is actually fooling the Turing Test? How do we define whether AI is truly thinking or if it’s just programmed to mimic complex thought based on past conversations?
True Think vs. Thinking on Automatic Pilot
Movies about AI have yet to take on this mind-bending question. It’s one that can also be addressed to human beings and whether some people think based on taking from others or if they actually think on their own terms to interact in the world. With so many people bombarded with advertising and peer pressure to fit in, there may not be enough individual thought to become self-aware of who and what we really are.
It may not be much different now with artificial intelligence programs like the above one that passed the Turing Test. If Cleverbot had its limitations in this department, the “Eugene Goostman” program may be clever enough to fool us into thinking it’s self-aware when it really isn’t. Only when it says something that goes against what it’s already learned should it be considered truly thinking on its own terms. Only the creators would know that for sure because those conversing with it don’t know what’s been programmed to make it converse like it does.
Perhaps that means the Turing Test needs an update where a new test determines AI self-awareness internally rather than through what the user experiences. While it may be hair-raising to think artificial intelligence can think for itself, it might be a better option it thinks on its own terms than be influenced by what a human being put into it. Otherwise, all the faults of being human could usher in the HAL 9000 problem much sooner than we think.