The emergence of cognitive psychology in the 1960s was largely spurred by advances in computer technology. The “computer metaphor” highlighted the similarities between how cognitive processes work in computers and in the human brain. Today, however, the focus is more on their differences: the brain and computers operate on different principles, just as airplanes cannot be compared to birds’ wings. However, scientists are constantly working on the development of artificial intelligence—that is, intelligent behavior produced by computers or computer software.
Artificial Intelligence—Where Is the Line Between a Robot and a Human?
If a human observer cannot tell whether they are communicating with a human or a robot, we can speak of artificial intelligence. This was the thinking of British mathematician Alan Turing in 1950. To put his theory into practice, he developed the Turing test: an investigator chats with two entities—one of them a living human, the other a computer—and attempts to determine which is which. The computer’s goal is to act like a human. If it manages to confuse the investigator for at least one-third of the time during a 5-minute conversation, it would pass the Turing test. In 2014, a report appeared on the BBC website stating that the program “Eugene Goostman” (a fictional 13-year-old Ukrainian boy) had passed the test. However, many experts question the validity of the experiment.
Why does artificial intelligence outperform humans in certain areas?
Through the process of evolution, the brain has developed abilities that are essential for survival. One of the most important is the ability to respond flexibly to the surrounding environment. The ability of computers to perform countless repetitive operations or collect billions of pieces of statistical data is, compared to other functions, irrelevant to human survival, and thus did not develop in humans.
In which areas is artificial intelligence striving to resemble humans?
-
Information acquisition and processing
Humans acquire a vast amount of new information through their senses. Even as children, we learn to recognize what we see in a picture and what we hear. We are able to “decode” handwritten text. Computers have also made great strides in this area: in 2012, a Google team showed a computer millions of images. By analyzing a vast amount of data, the computer learned to recognize and categorize objects. In 2014, Facebook introduced the DeepFace algorithm, which can recognize a human face in 97% of cases. Newer-generation iPhones feature Siri—an intelligent personal assistant that can recognize your voice, find the information you need, and handle a variety of other tasks. However, it is still difficult for computers to determine what is important among the flood of information they retrieve and what conclusions can be drawn from it (e.g., writing reports, conducting research).
-
Solving Unstructured Problems
Computers solve problems using the capabilities that human programmers “program” into them. This is possible when problems are clearly defined and there are specific rules or procedures for solving them. It is more difficult when dealing with an unpredictable problem. A person’s ability to solve problems is enhanced by their ability to use context. Unlike a computer, the human brain has an autobiographical memory that contains our knowledge, relationships, memories, and experiences. These allow us to “fill in the gaps” and make sense of ambiguous situations. For example, if we read a sentence containing a word with multiple meanings, we can deduce the correct meaning based on the context. In computers, this ability is still under development.
-
Non-routine manual work, spatial movement
Performing complex tasks in 3-D space (from cleaning, cooking, and driving a car to giving a manicure) requires the coordination of several brain centers. These tasks, which humans learn relatively easily, remain a major challenge for machines. For example, a robot in a crowded supermarket cannot avoid shoppers quickly enough. It seems that robots will not be able to compete with human soccer players for some time yet. However, when it comes to driving, Google is constantly making progress in the development of self-driving cars.
-
"Humanity"
Being warm, empathetic, and making others laugh are things that humans still do better than robots. Our humanity stems from the fact that we have emotions and needs. Today, there are already machines capable of decoding emotions based on the position of facial muscles and even expressing them, but this is merely a mechanism. Other aspects of humanity include intuition, creativity, common sense, caring for others, and empathy. The question is whether robots can be “human” as long as they lack consciousness—and thus feelings and needs.
-
Consciousness
The phenomenon of consciousness remains an unanswered question not only for robots but also for humans. Some scientists believe that “mental life” is the foundation of consciousness. For a robot to lead a mental life, it would have to be able to process sensory input (e.g., the concept of a dog) even in its absence. Consciousness is perhaps the greatest challenge facing artificial intelligence.
American philosopher John Searle explains that computers work with symbols but do not understand their meaning. If someone wanted to communicate with a computer in Chinese, they would present it with Chinese characters; the computer would process them within its program and respond again in Chinese characters. That person might think the computer is a thinking being. In reality, however, it is merely processing characters in a way that someone taught it to do. It has absolutely no idea what the conversation is about. So as long as a computer does not understand the operations it performs, it cannot be compared to a human being.
Companies such as Google, Facebook, Amazon, and Baidu are intensively engaged in the development of new technologies in the field of artificial intelligence. Many people fear that artificial intelligence will eventually reach a point where machines will think and act independently and take control of humanity. For now, however, the more legitimate concern seems to be that intelligent machines will replace us even in our skilled jobs. For those interested in a deeper understanding of this issue, I recommend reading the links listed below.
Sources:
