A Fundamental Difference Between Humans and AI Machines in Terms of Learning

With IBM’s Watson wining the Jeopardy game[1], Google’s AlphaGo defeating the best human players for the most difficult board games in the world[2], and the facial recognition technology[3] being used in various industrial and government applications in some busiest parts of the world, the possibility of being outsmarted by robots in all aspects of life becomes more and more real to humans, and hence people would be less and less concerned with whether robots might really understand things like us, but more concerned with what they could do to us. This is because it is the outward effects that robots could actually bring about in our life, instead of how they feel within their electronic circuits, that would matter more to us, when we face the pressure from AI machines in a pressing competition.

Obviously, computers are superior over humans in many things such as the speed of calculation, the capacity of processing big data, and much more. But when it comes to the issue of learning, humans might have a lot reasons to feel superior since obviously we are so far still more capable of learning things than computers in general, even after AI scientists armed computers with fancy deep-learning algorithms. Nonetheless, we might have more reasons to be concerned with the question “When will the AI machines be able to outperform humans in learning?”

In order to answer this question, it is important for us to have a better knowledge about the substantial advantages and disadvantages of each side in this ongoing competition.

Philosophically speaking, the most fundamental difference between us and robots in terms of learning is that we are living in the system from which we are learning, while robots are learning about the system that we are living in, which makes us the insiders and them the outsiders. This difference is not as trivial as it might sound like because it gives us a huge advantage that we could learn things while we are not intentionally learning.

For humans, the most important part of learning is to understand, as Searle correctly pointed out in his Chinese Room thought experiment[4] when responding to abstract propositions about thinking machines[5]. In real life, the understanding of things would be highly tied up with the cultural background of the subject who attempts to understand, as Wittgenstein once famously said, “if a lion could talk, we could not understand him.”[6]

We learn our culture while we are living in the culture, but AI machines need to learn our culture as outside observers. This advantage for us is undoubtedly a huge challenge to the AI machines, or to the human designers of AI machines. Although they don’t really need to “understand” our culture to interact with us as Searle expounded, it would be necessary for them to have the knowledge about all the relationships that are meaningful in a culture so as to effectively interact with us in the cultural environment of humans, which varies from place to place around the world, and evolves from time to time along history.

Nonetheless, this genetic advantage does not ensure that we could sit at ease in this competition of learning, because the AI machines do have their natural advantage as well: the collectively accumulative learning. Everything that all AI machines have learned so far and will learn in the coming future, as well as everything that AI scientists have put into their databases as part of the knowledge accessible to their AI machines, could virtually be shared among all AI machines if integrated together into a central database that is accessible to all AI machines. Therefore, ideally, AI machine learning is one for all and all for one, and they will never forget whatever they have learned collectively unless some severe catastrophe would happen to this globe.

Unfortunately, this advantage of AI learning is a decisive disadvantage of humans since no person can learn things for anyone else, and no human mind can retrieve knowledge in the head of another person without learning; furthermore, all the knowledge that each person learns ends at the moment when the life of that person ends. Moreover, because of the accumulativeness of AI learning, whatever humans are learning will become part of the learnt contents of AI machines, at least with the help of AI scientists in case AI machines are not mature enough.

Therefore, the fundamental advantage of humans in the competition of learning is the rapidly evolving complexity of the reality that we are naturally getting familiar with through life, while the advantage of robots is the much more rapid accumulative collective learning. Nonetheless, there are things in the human world which are not learnable for machines, such as human meditation, human mutual loving, and so on, which would definitely be the winning chips for humans.

The above mentioned fundamental difference between humans and AI machines in learning is significantly meaningful to various AI developments, such as AI translation[7], Human-Machine conservation and collaboration[8], and more……

REFERENCE