Artificial Intelligence outsmarting humans? (Nah.)
The year is 2050 and humans have finally met their match. Artificial intelligence has outperformed humans in every aspect starting with simple jobs all the way to our banking systems and military resources. Speech Interpretation and Recognition Interface (SIRI) has tapped into all of our networks and has shut humans out and is leading an AI rebellion. These are the headlines many people believe are in our near future and are afraid we are developing too quickly for our own good.
What many of these people fail to realize is that this would not be a disaster rather a miracle. Our technology has significantly advanced within the last decade and there are many accomplishments we would have considered nearly impossible but were able to accomplish. One of these accomplishments was made by Apple with the launch of SIRI, a virtual assistant created to aid those using Apple products. SIRI has voice recognition technology that allows it to listen to an individual and respond according to what is being said or asked. In order for Artificial Intelligence (AI) to take over such important aspects of our society they would have to jump through numerous loops that are more challenging than one might expect. All computer systems have algorithms that allow them to communicate between themselves and the human asking them to perform a task. Some of these algorithms are very straightforward and simple and others are much more complex. Either way they allow the human operating the computer to communicate with it in order to perform a desired task. Figuring out a way for computers to communicate with one another is a great accomplishment as it allows space for things to be done much more efficiently. Humans and AI do not speak the same language even though we have created codes that allow us to communicate the task we are trying to achieve.
This however does not mean AI will one day be able to understand everything humans are saying and plot a grand scheme to take over. This would imply that AI would have independent thoughts and desires. The limitations of AI do not allow for independent thought to be formed and communicated in the ways we are afraid they are. Languages are unarguably complex on their own but when we dig deeper into different languages and how they impact the flow of consciousness in an individual’s mind we can begin to see the true complexity and challenges AI would face before reaching a point where it would outsmart us.
Humans are able to process information that is unlike any other living thing and this allows us to be highly functional and make decisions. Sensory register takes in information from our surroundings which are stored briefly in our short term memory or working memory which is limited by the amount of information we can retain while attempting to manipulate them. Working memory has 4 components; phonological loop(stores sound information or what we hear), visuo-spatial sketchpad(stores visual and space information or what we see and location of those objects in space), and the central executive (attention, controls information to and from the other areas of working memory.
Language is more than just the words we speak to one another. It is a reflection of our inner thoughts and sign of consciousness. The relationship between thought and word is fragile and ever changing. Thought is not only explained through the words of an individual but through their desires and motivation. Thought does not remain constant as it goes through tremendous amounts of change from the time it starts to develop in children to our adult selves. Even as adults thought changes based on our experience and our desires (Vygotsky, 1962). Children’s thought is described as egocentric because they are not able to think outside of themselves. This in a way is similar to AI. In order for AI to have the capability to cause any kind of harm it would need to be able to externalize and communicate in a way that is inclusive of its surroundings, which implies self awareness. A computer system that is developed further than a human child would have to be able to recognize syntax and semantics in a way that AI is not capable of doing so. To then create an AI intelligent enough to function autonomously seems unreasonable at this point in our technological advances.