How can we implement the reason and emotion needed to develop AI?

The development of strong artificial intelligence requires technologies that embody human-like reason and emotion, which can bring great benefits to human society.

 

From driverless cars, AlphaGo, fingerprint recognition devices, to robots that think and act on their own that we see in movies, they are all machines based on artificial intelligence. However, robots that think and act on their own are far different from our current artificial intelligence technology. So, what does it take to create a robot that can think and act on its own from a simple machine?
First, let’s take a look at the difference between weak and strong artificial intelligence. Fingerprint recognition machines and driverless cars are classified as weak AI. They don’t have a sense of self and can only do what humans program them to do. AlphaGo, which we all know, is one of them. There are two types of AI: general and generalized. A generalized AI simply specializes in a specific field, but a generalized AI can learn to cook something it has never done before by watching someone else do it, or it can take over a simple task and perform it. General-purpose AI can be thought of as a step from weak AI to strong AI.
The Terminator, the robot in “iRobot,” and Jarvis in “Iron Man” are examples of strong AI that act with self-consciousness. The main difference between strong AI and weak AI is the presence or absence of an ego. If it has an ego, it can think, judge, and act on its own. To do this, it needs a level of reason and emotion similar to humans. Not everyone would question the need for reason in AI. Reason is something that is lacking in current AI technologies, but it is present to some extent. Making judgments based on a large amount of data can also be considered reason. However, many people question whether emotions are necessary for AI. Some argue that giving A.I. emotions will cause problems. However, reason and emotion are not as dichotomous as many people think. Emotions are necessary for reason to be fully expressed, and therefore, in order to develop strong AI, we will need a much better way to instill reason and emotions than weak AI.
There is a lot of debate about whether we should develop AI. However, there is no doubt that there are many benefits to human society. We can solve these problems if we prepare for the problems that will arise and start developing them. So, let’s see how we can develop reason and emotion to create AI. In order to move from weak AI to strong AI, we must first develop the reason of weak AI to a much higher level. Weak AI relies on expert systems to process huge amounts of data. Expert systems are trained by case-by-case learning, which means that they learn to recognize different outcomes in different situations depending on the conditions. However, this system is simply a system that processes a lot of data quickly, so it can only be solved by increasing the power of supercomputers. This is where machine learning comes in. Machine learning is a technology that allows computers to create rules on their own to analyze new data, unlike expert systems that only analyze the data they are given. Machine learning allows computers to have regularity, which is a similar trait to humans.
Another technique for giving machines reason is genetic algorithms. Algorithms as we know them are used to solve math problems and are very useful for coming up with clear answers. They can handle many math problems automatically, and they can leave complex calculations to computers to produce accurate values. Unlike math algorithms, genetic algorithms take into account mutations and inbreeding in a population to find the most appropriate answer over generations. There are many variables in human life, and there is no single answer. It’s all about finding the most appropriate answer, and AI has this quality that allows it to make decisions that are more in line with human judgment.
For weak AIs to become strong AIs, they also need emotions, but it’s not yet clear how to give them emotions. While it is not yet technically possible to create subjective emotions, it is possible to program some emotions and act accordingly. For example, you can program a robot to feel sadness at a funeral, or joy at a birthday or anniversary. By entering so many situations, you can expand the range of emotions and express them through reason. The fundamental problem with this approach, however, is that it doesn’t give you the emotions themselves, but rather mimics them with reason. This approach will never be able to fully express emotions, and will ultimately be limited in its ability to be human-like. A corollary to this is that different people feel different emotions in the same situation. On the same birthday, one person might feel joy, while another might feel lonely. This diversity of emotions needs to be implemented, so machine learning or genetic algorithms should be applied not only to rationality but also to emotions. Furthermore, if we conduct research on the role of emotions and how they work, and create a model that mimics them, we will know how to give emotions to strong AI.
If we give reason and emotion to weak AIs in the way suggested above, we will be able to create strong AIs. With further advances in technology, strong AI may not be a distant dream.