In Seoul, South Korea, a Google-created artificial intelligence has been squaring off against a mortal man in the 2,500-year-old strategy game, called Go, that’s several orders of magnitude more complicated than chess.
When it was finally over, Google’s AlphaGo won four out of five matchups, making AlphaGo a role model for young artificial intelligences everywhere.
How did the robot pull off such a decisive win?
Wired reported that “AlphaGo relies on deep neural networks—networks of hardware and software that mimic the web of neurons in the human brain. With these neural nets, it can learn tasks by analyzing massive amounts of digital data.”
That’s bad news for SEOs the world over, because Google isn’t just using neural nets to beat Koreans at board games, it’s also using these advanced networks to make their search results more efficient. And in the process Google might just create the artificial intelligence science-fiction authors have been dreaming about for decades.
When will Google become sentient?
For eons, our foremost philosophers and scientists believed that man’s ability to make tools set him apart from the animals. It was actually Jane Goodall who smashed this macho theory to bits after she witnessed chimpanzees making rudimentary tools.
Now, many scientists instead believe that it’s our unique capacity for language that truly sets us apart from the beasts of the world. In the quest to give birth to artificial intelligence, computers’ inability to comprehend human speech is one of the main stumbling blocks – as anyone who’s ever tried to talk to Siri can attest.
Yet we’re getting closer to that Brave New World every day, and in many ways Google is leading the charge. Not only does the company control one of the most complex computer networks on the face of the planet, but by necessity Google is making major strides in Natural Language Processing.
Because search engines can’t (yet) understand a piece of content or search query the way a human does, the company uses advanced algorithms to analyze the semantic content of content. Recently, Google started using neural net technology to push its NLP efforts forward.
And while Google is using machine learning to bring you more relevant search results, and Apple is doing the same to make Siri more useful, both tech companies are striking at the very core of human consciousness.
How long until Siri can pass the Turing Test?
This March it seemed like every website on the internet covered the story about Siri and Cortana responding to questions about violence and suicide. As Gizmodo summed it up, “Siri Is Woefully Ill-Equipped to Help With Your Mental Health Problems.”
Clearly we haven’t quite cracked the secret to artificial intelligence, even despite recent breakthroughs in machine learning (see my earlier post on Google’s RankBrain).
So why does it matter that Siri, Cortana, and Google Now still struggle to understand even basic questions?
Alan Turing, the father of the computer, is known for many things, but AI geeks remember him best for the so-called Turing Test. The test has evolved through a number of different versions over the years, but simply put, the Turing Test is a method for determining whether a computer can think like a human being.
Per the Webopedia version of the experiment…
“The test is simple: a human interrogator is isolated and given the task of distinguishing between a human and a computer based on their replies to questions that the interrogator poses. After a series of tests are performed, the interrogator attempts to determine which subject is human and which is an artificial intelligence.”
If a computer can genuinely pass the Turing Test and convince a human that it’s a fellow living, breathing, talking person, then that machine can fairly be called Artificial Intelligence, with a capital AI.
One day, when you pull out your smartphone to ask Google or Siri a question, instead of returning a list of search results, they just might answer back.