Types of AI

     How many types of AI do you think there are? The answer to this varies by what kind of AI we are talking about. As Kaplan and Haenlein of ESCP Europe states, "AI is still a surprisingly fuzzy concept and a lot of questions surrounding it is still open." From my research to this question as well, I found out that there is not only one answer to how many types of AI there are but numerous. In this blog post, I would like to share two interpretations and give my thoughts on it.

    First, Kaplan and Haenlein states that there are three types (or stages) in AI: analytical AI, human-inspired AI, and humanized AI. They borrowed studies that investigated the skills specifically regarding to business and looked at the patterns shared by managers and employees who performed above average. For this evaluation, they used four criteria or skills to categorize AI: cognitive intelligence, emotional intelligence, social intelligence, and artistic creativity. Cognitive intelligence is about how AI recognizes patterns and thinks systematically. Emotional intelligence relates to how AI adapts, acts to achieve their goals, has somewhat emotional self-awareness, and so on. Social intelligence is about how empathetic an AI is, if they can work in teams, and can be an inspirational leader. Artistic creativity asks if AI can create artistic works such as paint or play an instrument.

    Based on the four criteria, there are five stages in performance. Below the three AI types, we have expert systems, which are "collections of rules programmed by humans in form of if-then statements". They are not considered AI because they do not have the ability to learn independently from external data without the help of humans. Expert systems do not have any of the four skills. Analytical AI only has cognitive intelligence. Examples include image recognition or self-driving cars. Human inspired AI has cognitive and emotional intelligence. Most of the AI used in companies today are human inspired AI. These include that can understand human emotions and use that to make decisions. Humanized AI has cognitive, emotional, and social intelligence. Although these types of AI are still in-progress, these systems will have the ability to be self conscious and aware in their interactions with other people. We can see that artistic creativity is a skill that is very hard for AI to mimic, as if AI can copy a certain art, it will not be original and never surpass the true form of human imperfection. 

    I was able to agree with Kaplan and Haenlein's categorization. It is simple and very easy to understand. It shows that there are three stages in AI and that each step brings a big change into what the AI can do and not do.  

    The other analysis I found put AI into four types: reactive machines, limited memory, theory of mind, and self awareness. Arend Hintze of Michigan State University focused on machine learning and deep learning. In his perspective, these still do not have the skills to build the machines we think we will have in the future with their own consciousness. 

    Reactive machines are the most basic types of AI. They do not have feelings and cannot use their past experiences to make decisions because they do not memorize them. Limited memory is a type that has the ability to look back into their past. Self-driving cars is a good example. They use their cameras to observe cars' speed and direction for a period of time to make their decisions. This type is called this way because the AI's memory is limited. The program does not have the skill to stock up past experiences like we do in order for them to use them when needed. The memory they analyze and keep will only be used for that short time period. Theory of mind is the type where they will be able to recognize other entities in this world. They are not fully made yet, but will have the ability to understand that living species "can have thoughts and emotions that affect their own behavior." Self-awareness will be AI that can form representations about themselves, Machines will not only have the skill to understand consciousness, but have their own consciousness as well. Self-awareness is an extension of theory of mind. Hintze states that we have to understand the human's memory, learning, and the ability to use past experiences for current decisions in order to create machines that have the ability to do that. 

    I believe this is a similar approach to classifying the types of AI as the one Kaplan and Haenlein did, but including two types of AI that has not been made yet. Should we count the types of AI that do not exist right now but we think are bound to exist in the future? I think that putting some vision into the possibilities of AI is important, but believe that not enough is understood on what kind of AI we have right now. Both analysis deeply involves emotions and memory. I never came to think deeply about how special our emotions and memories are. They are very complex and even if AI can imitate it, it is hard for them to use that knowledge for future decisions. Although artistic skills are an important factor in determining an AI type, memory and emotions will be the core of understanding and building AI, as memory made by emotions is what drives humans into their very own actions. 

This was our blog to question #2 of our 30 research questions.

References:

Hintze, A. (2016, November 14). Understanding the four types of AI, from reactive robots to self-aware beings. The Conversation. Retrieved July 2, 2022, from https://theconversation.com/understanding-the-four-types-of-ai-from-reactive-robots-to-self-aware-beings-67616

Kaplan, A., & Haenlein, M. (2019). Siri, Siri, in my hand: Who’s the fairest in the land? On the interpretations, illustrations, and implications of artificial intelligence. Business Horizons, 62(1), 15–17. https://doi.org/10.1016/j.bushor.2018.08.004