What is Artificial Intelligence?
Last week we discussed what the differences are between artificial intelligence and machine learning. The next couple weeks we are going to do a deep dive into AI and ML. This week we will start with AI.
Alan Turing, the man who played a crucial role in cracking the Enigma machine during World War II, later published an article entitled “Computing Machinery and Intelligence” asking the question, “Can machines think?” thus creating the Turing Test. The Turing Test is a method where a human tries to distinguish between the responses from both a human and a computer.1 Artificial intelligence can be described as the branch of computer science trying to answer Turing’s question by ongoing attempts to get machines to replicate the human mind when it comes to decision making and problem solving.2
Stuart Russel and Peter Norvig created a textbook entitled Artificial Intelligence: A Modern Approach that became one of the leading textbooks in AI studies. They explore four different approaches to AI2:
Thinking Humanly
Thinking Rationally
Acting Humanly
Acting Rationally
Thinking humanly refers to the idea of the computer trying to understand and mimic how the human mind works, whereas the “thinking rationally” aspect uses logic to make decisions. Acting humanly, is related to the Turing Test approach in that it answers the question of “Can this computer pass as a human?” Acting rationally would be the computer behaving in ways to achieve end goals3.
There are four types of artificial intelligence: reactive machines, limited memory, theory of mind and self-awareness.
A reactive machine is the most basic of AI systems. It cannot form memories or use past experiences to make active decisions. These kind of machines “perceive the world directly” and are only designed to act on what is seen and perform specialized duties4. An example of a reactive machine would be Deep Blue, IBM’s chess playing computer. Deep Blue was limited to the knowledge of identifying the pieces on the chess board and knowing how each piece moves without any regard to previous moves or future plays. Deep Blue went on to beat international grandmaster Garry Kasparov on that knowledge alone4.
Limited Memory refers to AI that can store previous data and weigh future decisions. Self-driving cars would be an example, in that they observe other vehicles’ direction and speed and monitor them over time. However, the vehicle does not learn from these experiences the way a human would and become a better driver over time.
Theory of Mind is simply the theoretical future of machines. While we do not yet have the technology and scientific capabilities to reach this level of AI, the premise is that machines would learn other living have thoughts and emotions that affect their behavior, therefore they would also understand the concept of making decisions based on their own self-reflection and adjust accordingly2.
Finally, self-awareness, which is the final step once the Theory of Mind is established. In other words, the machines start having their very own consciousness2.
Overall, artificial intelligence is still very much in its infancy. As with Theory of Mind, we have no way of knowing if and when we will we even have machines capable of their own consciousness. The human mind, however, has no limits to its imagination. We’re just getting started.
By: IBM Cloud Education. “What Is Artificial Intelligence (AI)?” IBM, 30 June 2020, https://www.ibm.com/cloud/learn/what-is-artificial-intelligence.
“Artificial Intelligence.” BuiltIn, https://builtin.com/artificial-intelligence.
What Is the Difference between Thinking Rationally and Thinking Humanly?, https://philosophy-question.com/library/lecture/read/43022-what-is-the-difference-between-thinking-rationally-and-thinking-humanly.
Hintze, Arend. “Understanding the Four Types of Artificial Intelligence.” GovTech, GovTech, 14 Nov. 2016, https://www.govtech.com/computing/understanding-the-four-types-of-artificial-intelligence.html.
