Artificial intelligence (AI) is the ability of a computer system to perform functions that generally require intelligent thought. Although the system can make choices and determinations about problems posed to it, it can only respond to the extent it’s programming allows. The computer systems inability to think outside this programming ultimately limits its usefulness. What exactly is the limiting factor?
According to the Business Encyclopedia, the intent of AI research is to develop computer systems having the ability to perceive and to learn, to accomplish physical tasks, and to emulate human decision making.” Furthermore, researchers attempt to design and develop intelligent systems as well as understand them. Although there has been significant advancement in the field of AI, such as neural networks and expert systems, the key factor restricting AI evolution is the lack of commonsense knowledge.
Researcher Philip Eni-Dor, in writing Artificial Intelligence-A Short History and the Next 40 Years, explains that commonsense knowledge is that knowledge everyone should possess and
is essential to interpreting the information acquired by the senses and in understanding natural language, which itself consists basically of real-world phenomena, sensory stimuli, and behaviors.
Understanding natural language requires the ability to process speech. Speech processing is all about input and output, and each is accomplished with varying levels of difficulty. Speech synthesis (output) is already well developed and is used extensively across the computing world. However, speech recognition (input) is limited in use and much more difficult to master because understanding speech requires an understanding of context, which requires cognition. Furthermore, interpersonal differences in speech patterns are greater than individual differences. Therefore, I believe the future of AI hinges on the ability of computers to understand human speech. The ability to understand speech at the phoneme-level and to differentiate between nuances such as voice inflection is the key to obtaining commonsense knowledge.
To understand how speech recognition relates to commonsense knowledge, consider the following statement: “I took the money.” This is a simple admission of guilt, but if one of the words is stressed, the statement actually becomes a question and an implication:
- “I took the money?” says I didn’t take the money, but I may know who did.
- “I took the money?” says I didn’t take the money, but I may have done something with it.
- “I took the money“ says I didn’t take the money, but I may have taken something else.
As one can see, a computer system’s ability to differentiate between voice inflection, tone and pitch would provide significant insight into understanding human emotion. In fact, there’s software in use in call centers today that will automatically alert supervisors when a customer’s voice pattern indicates frustration or anger. The software actually understands the customer’s behavior and acts accordingly. This will only improve as time advances.
Artificial intelligence is the imitation of cognitive (human) thought by a computer system. Artificial intelligence systems certainly help humans better understand data relationships; however, this helpfulness is limited because the computer system can not think past its programming. The ability to understand human speech could change this and would possibly allow the computer system to actually communicate on a cognitive level with a human being. The implication of this achievement is the word “artificial” may no longer accurately describe the computer system. It might just be considered “intelligent”.

The new theme is called 



Recent Comments