A Reflection on Human Intelligence and Artificial Intelligence
The famous “Halting problem” proposed by Alan Turing shows that we will never know if a problem is unsolvable by computers because of the binary nature of the question. There can only be two answers, either “solvable” or “unsolvable”, there is nothing in between. Similar to the famous Schrodinger’s cat experiment, where we will not be able to tell if a cat is alive or not without opening the chamber, we cannot know if a loop is infinite without breaking out the loop (that is “our curiosity kills the problem-solving process, not the poor cat this time”).
Arising from the “Halting problem”, I hereby propose another problem in the field of artificial intelligence that resembles the dilemma faced by Turing and Schrodinger. The age of information has brought us unbounded possibilities. Data, which infiltrates every aspect of our life is so informative that we can use the results from mega data analysis to conduct market trend prediction, customer preference classification, material behavior improvement, molecular interaction assessment, and many many other fields. Further, with the rise of artificial intelligence (A.I.), machine learning (M.L.), and deep learning, the decision-making process is becoming unprecedentedly efficient. But, are the techniques and algorithms really “intelligent” after all?
So much for the foreshadowing. The perspective that I want to present in this article is “Can we ever know if artificial intelligence is intelligent?”. It is true that deep learning algorithms can mimic the human brain, the intricate neural connections. However, what will happen when intelligence is no longer “artificial”. The word “artificial” arise from “being created by human” (trust me, I actually looked it up on the dictionary). When artificial intelligence become so advanced, when it can deal with problems that have so many dimensions that is incomprehensible by human intelligence, then, will we be able to detect its superiority?
The dichotomy of being intelligence and unintelligent is human-defined, driven by the “limited” power of the human brain and pursuit for a replacement for human intelligence. When artificial intelligence indeed reach the level that humans cannot interpret (i.e. when it exceeds the the scope that is defined by human), then we will never say that it is intelligent.
Not being pessimistic about human intelligence, the previous logic has a condition: human brain is static. However, we cannot rule out the possibility that the human brain improves alongside with the advancement in artificial intelligence. Human brain, exhibiting fascinating neural plasticity, can adapt to the changing environment. The plasticity is what distinguish us human from other beings. Thus, when we are enjoying the convenience brought by artificial intelligence, let us stay open-minded for new possibilities and always stay hungry for knowledge.
2020.12.29