Known as one of the best methods of testing for artificial intelligence, Alan Turing’s Turing Test may soon be capable of being passed by artificial intelligence. The Turing Test is a fairly simple concept: If a judge cannot reliably tell a machine from a human during a normal conversation, then the machine is said to have passed the test.
While currently today there are no machines out there that have been accurately able to pass the test, we are coming alarming close to a day when this occurs. Consider such popular technology that everyone has today. Machines like GPS systems or Apple’s Siri are now able to have full conversations with a human and can give fairly accurate responses most of the time.
While these machines cannot currently pass the Turing Test, cognitive scientist Robert French recently wrote:
“Two revolutionary advances in information technology may bring the Turing test out of retirement. The first is the ready availability of vast amounts of raw data — from video feeds to complete sound environments, and from casual conversations to technical documents on every conceivable subject. The second is the advent of sophisticated techniques for collecting, organizing, and processing this rich collection of data.”
If you think about this concept for a second, French does have an entirely valid point. Think for a second how much information has been gathered in these past few years from video cameras, smart phones, GPS systems, and even satellites. There is an insane amount of data out there that has been captured that seems like it could be implemented into a system to potentially fool a human being.
These advancements are amazingly impressive but they’re not exactly where 20th century scientists thought we would be at this moment in time. At one time, the human mind was thought to run very similarly to that of a computer, where our brains run logical commands similarly to that of a computer. During the mid-20th century, many scientists believed that within a few decades, humans would be unable to be able to tell another human and a computer apart from each other.
Unfortunately, this idea is way more complicated than many scientists thought and appearing human has been met with tons of different challenges. Things like logic cannot simply be programmed into a system very easily.
Although there have been many famous systems out there like the Watson computer, which famously defeated Ken Jennings, the highest-ranking human Jeopardy player, these types of systems only seem to be demonstrating a tiny bit of what occurs during human thought. The thing is, you can ask Watson a question that he will most likely be able to answer but if you tell him something like “Write me a poem,” the computer simply cannot do it. The scope of these systems is very limited in terms of what they can actually do.
While this whole theory of data aggregation possibly being the key to defeating the Turing Test and possibly becoming a major breakthrough for artificial intelligence, many people like Satinder Singh, an artificial intelligence expert who works at the University of Michigan, are somewhat skeptical. “Are large volumes of data going to be the source of building a flexibly competent intelligence? Maybe they will be,” he said.
It is important to note, however, that pulling together all of this data and figuring out which parts are important to have a computerized system remember is a very difficult process. If you really think about it, what are the best pieces of information to remember? And how would you be able to program a computer to think logically on its own in the same way a human brain does? This type of problem clashes with the major questions about ourselves as humans when we ask “Why do we do what we do?”
No matter what the outcome is in the next few years, we should be seeing even more amazing innovations relating to artificial intelligence in the near future.