Ray Kurzwell is 65 years old and an accomplished inventor building anything from music synthesizers to speech recognition systems. He has always had the dream of creating an "artificial intelligence" that would rival human intelligence and he doesn't think that's too far off. Since he has started working at Google he is hoping that the resources the company has will help him draw closer to that dream. Can "Google Search" evolve into an electronic friend as an "AI" that can recognize what you need ahead of time and converse with you like a normal human being?
Presently, many searches on Google (GOOG) are done in the form of questions. Imagine the search engine having a better understanding of the natural language with the ability to do a better job answering those questions by understanding what web-pages are communicating. It would give the "search" the ability to think for itself, understanding the ambiguities and common sense that comes with speech. The closest example of this right now would be a computer by IBM called Watson. The computer actually read Wikipedia and a few other dictionaries before it played "Jeopardy" on TV.
Mr. Kurzwell has access to something at Google where he doesn't anywhere else: Google's vast data and computing infrastructure. These are the two things needed to reach the next level of AI. There is no better person or better company to mix together than these two to create this next level search engine. Google has its "Knowledge Graph" with something like 700 million topics that includes people, concepts, places ETC. this will be the knowledge base of the new search. The potential of what these two can do together is mind-boggling! Imagine a person being born. This person, as an intelligent being, takes in knowledge and bit by bit creates patterns of understanding and this is how we learn. Imagine an artificial intelligence as a search engine that is able to learn how to assist you the same way you learn how to develop a relationship with somebody. This intelligence understands more and more as it interacts with you what you are looking for and what you need.
The key to building a system like this is access to knowledge and "semantic" meaning. If a system can be developed that has access to these hundreds of millions of web-pages, then a search could be pinpointed perfectly for us. Can you imagine how quickly knowledge could be gathered, studies could be done, or research could be put together?
To get to this point is a formidable challenge. Understanding the concept of "semantic meaning" across different cultural oceans is the first challenge. How our mind processes and categorizes things is incredible and on the scale that Google is attempting to build the same process amazes me. Imagine taking a phrase and then building a library of all the different ways something could be said. Then translate and combine that with a learning algorithm so that the search engine not only understands but can also learn from you-- how you talk. The evolution of language will make this an ongoing project that will never have an end.
I cannot say how long it will take before we have a search engine that can understand us and know exactly what were looking for. To build something like this that simulates how our brain works will take years because first the artificial brain has to be built, studied, and then researchers can identify methods and design algorithms to create the AI. Once it is created, technology will accelerate its learning abilities.
How Will this Affect Google Stock?
Google is already the leading search engine, this is one of the reasons the stock continues to grow so well. In the future if people can find what they were looking for faster because of a "Google AI" then they will continue to go to the search engine before anywhere else. This means high traffic which translates into advertising revenue for the company. So when I write about the possibility of a new type of search engine that Google creates, it will translate into better revenue streams for the company through advertising. If other engines cannot keep up, they will fade away. When I can look for something through a "Google AI" and find it much more accurately and quicker than anywhere else, I will be prone to go there every time.