At the very first place, we will have to admit the truth that we are not winning over robots while completing a specific task. This is inevitable as they can almost unlimitedly upgrade themselves, while practicing when we are all sleeping. Yet, does this mean we humans are losing all the time?
Imagination of AI
We often hear gurus say that robots lack the ability to imagine. Indeed, even artificial intelligence is manipulated by the command set and programs. AI only has the ability to partially change their working mechanism, like leveraging neural network to improve its performance, which cannot be done by traditional programs. Yet, AlphaGo still cannot turn itself into a program that is able to do things other than playing Go. Therefore, something fundamental is still kept constant and cannot be edited by AI itself. Their potential seems to be limited by its designed purpose. An easy conclusion seems to be drawn: AI just can’t imagine like us. But is this statement true?
From a philosophical perspective, what we thought to be imagined by us is in fact just a mix of existed concepts. In other words, imagination isn’t creating something from zero. This contradicts with our common cognition, as we often define “imagining” as an activity that is able to create things that are completely new. For instance, we all haven’t seen a unicorn. So unicorn seems to be a product of human imagination. Yet, if we think deeply, a unicorn is just a mix of a body of a horse, horn and perhaps a pair of wings. We might not have a chance to see a unicorn, but we do have seen the three components that make up our imagined animal. The same mindset can be applied to everything that we have created. And as long as something is generated according to a regular pattern, it implies a chance that AI can learn it through practicing.
We can imagine that a future machine with much more enhanced processing power is pairing up thousands of object, while thinking about the effect of merging a set of things together to become a new product. The working principle will be letting the AI system to have a comprehensive understanding of a basket of objects, including the physical properties, designed purpose of the object etc. The system then automatically starts calculating the probability of combining several parts to achieve a particular goal. Imagination power can therefore be possessed by AI, or at least, being imitated by AI.
This shocking result can be disastrous to us, as people that seem to be unable to be replaced by AI, like creative work producers, will now face unemployment. In fact, we are already experiencing the power of so called creative programs. Google, being the pioneer in AI, has demonstrated the ability to write a mournful poetry solely by their own program. This proves that AI can indeed do something that most of us thought they cannot.
Don’t Be Too Worry Yet
What I have stated will not be something that is easy to be done. Personally, I think such a powerful program that is able to perfectly create a certain product to tackle our problem needs more than 10 years to be created. After all, AI intellectuals will have to design algorithms that enable the program to dissect plentiful inventions to understand how they work. It will take quite a long time for training the program.
Second, we have to believe that every technological revolution won’t just destroy the original world, as it simultaneously creates things that are beyond most people’s imagination. For example, the emergence of steam-powered machines caused panic across workers. Yet, restricted by the background of era, most of the people at that time did not realize the massive opportunities beneath. New machines required more skillful workers with specific techniques to operate them. Hence, like those who lived in the past, we are limited by the current situation, making us extremely skeptical about AI systems. What we need to do is to think about the treasure brought by the upcoming AI revolution, so as to embrace benefits of this great invention. The key is to equip ourselves, not blaming the world. That is what keeps us “untouchable” as mentioned by Thomas Friedman.
For the very last part, we have to answer the question set at the beginning. I think AI will eventually replace all jobs that we have seen, from artists to politicians. However, like what has been mentioned, new opportunities will again be created. So, to be more precise, AI will only partially replace humans, but not the whole part of us.