The onslaught of Artificial Intelligence in every domain of human life has raised a very pertinent question: what are the aspects of human intelligence that cannot be replicated by Artificial Intelligence?
The answer to the above question would also provide the direction for the development of human capital in an economy dominated by Artificial Intelligence.
What is the essence of human intelligence? In the 16th century, French philosopher Rene Descartes, while writing his book, “Discourse on the Method”, gave the first principle of his philosophy: Cogito Ergo Sum. I think, therefore, I am. This proved that having a thought about oneself proves the existence of oneself. The capacity to think about the very act of thinking makes a human being a thinking being.
Artificial Intelligence aims to simulate human cognitive activity for problem-solving and decision-making. The question is, can Artificial Intelligence think? Before exploring the answer to the question, let’s discuss how thinking is defined.
The answer to “Can AI think?” depends upon how we would define thinking.
Not everything that comes to our mind is thinking. Daydreaming is not thinking, and neither is intuition or gut feeling. Thinking is defined as a cognitive function used in the process of making choices and judgments. It has a direction, purpose and process. In the thinking process, a person arrives at a conclusion through reasoning. Thinking is essentially question- or problem-driven. The high-order thinking comprises reflective thinking and critical thinking.
AI is good at solving “decidable” problems that can be solved by constructing a corresponding algorithm to answer the problem correctly. These problems can be solved mechanistically by following a set step-to-step procedure. The ability of AI to solve decidable problems is not based on “Deep Thinking” but on “Deep Learning”. The distinction is important as Deep Thinking comprises higher-order thinking skills such as reflective thinking and critical thinking. Deep Learning is the ability of machines to learn by example. It does a lot of number crunching in a smart way to manipulate large data sets. The forte of human beings is Deep Thinking which cannot be imitated by AI. As Cal Newport, MIT-trained Computer Science Professor, has written: the ability to solve undecidable problems would be increasingly valuable in the economy where AI would be taking over repetitive tasks using technologies such as RPA (robotic process automation).
He coined the term Deep Work to come up with the solution of undecidable tasks. Deep Work is defined as professional activities performed in a distraction-free concentration that pushes your cognitive capabilities to their limit. These efforts create new value, improve your skills, and are hard to replicate.
Therefore, human beings are required to perform high-value tasks that require creativity, ingenuity and strategic thinking. Hence, the ability to Deep Work is the human powerhouse that is out of the reach of AI.
Another aspect of human capability that remains elusive to AI is the understanding of the context. Context is everything. The AI has access to a vast trove of data, but humans are exposed to a diversity of contexts. Artificial Intelligence lacks critical thinking processes or judgment calls as they are applied to varied contexts. Human intelligence weighs the recommendations, recognizes patterns from past experience, and makes decisions based on understanding the context.
To understand something, it is important to know what it is and why it is the way it is. Through the manipulation and processing of information, AI cannot arrive at the human level of understanding. As Nobel Laureate in Physics, Richard Feynman told how his father taught him the difference between knowing the name of something and knowing the thing. Once, he told him, pointing to a bird, it’s a brown-throated thrush and then gave the names of the bird in different languages. Then he taught him the lesson that even if you know all those names of the bird in different languages, you still know nothing about the bird – you only know something about people, what they call that bird.
Chiding the Google Brain Project in which 16000 processors took a few months to identify the “shape of a cat” from vast data of stills from video clips, Piero Scaruffi, a widely acclaimed author in the field of cognitive science and artificial intelligence, asked how much time a kitten would take to recognize a cat with its still underdeveloped brain: a split second. He emphasized the point that human intelligence is something else as humans not only do the fairly easy task of recognizing that a cat is a cat without seeing millions of images of cats but also know the whole story behind cats. We know what cats do and can relate to, for example, Schrodinger’s Cat, and that’s deep thinking, and machines don’t do that.
Empathy is another skill that would be crucial to possess in an era when machines would act like humans and humans would act like machines, as warned by social psychologist and philosopher Erich Fromm. AI can recognize a person’s face through a face recognition attendance machine but would never understand that person, and understanding would remain an exclusively human domain. The human touch cannot be replaced by AI touch. However, the dependency of humans on virtual assistants like Siri and other networking tools would make face-to-face interactions less frequent, thus diminishing our capacity for empathy.
Sherry Turkle, Professor of the Social Studies of Science and Technology at the Massachusetts Institute of Technology, asked the question, Is AI killing empathy? “With face-to-face interaction, we learn the skills of negotiation, of reading each other’s emotion, of having to face the complexity of confrontation, dealing with complex emotion”, she argues while calling herself pro-conversation rather than anti-technology. To develop empathy, we need spaces for conversation where we can disagree on ideas while developing an understanding of others’ points of view. The skill of empathy, of relating to others, would become essential in the time of the proliferation of automated systems.
The applications of Artificial Intelligence are bound to grow and would help human beings in solving problems across all fields, but if we lose our focus on capitalizing on the power of the human brain and its cognitive capabilities, we are in danger of abdicating the essence of human intelligence. It would suffice to remember that AI is masterful in Deep Learning but inept in Deep Thinking. AI would be excellent in knowing the contents but poor in understanding the context. And finally, AI can imitate to be empathetic but lacks authenticity.
Artificial Intelligence can be no match to human intelligence, only if we do not lose being human.