As a curious and researcher in the field of artificial intelligence, I see quite interesting comparisons about human intelligence and artificial intelligence. One of the most used are variations of the book “Robot: Mere Machine to Transcendent Mind”, from 1998, by Hans Moravec. Moravec is from Austria and has been active in the United States for a long time. He is, among other things, a futurist, a name given to people who are dedicated to future studies. His publications and predictions are focused on the topic of transhumanism. Transhumanism, in turn, is a philosophy that aims to improve the human condition through the use of science and technology (biotechnology, nanotechnology and neurotechnology) to increase cognitive capacity and overcome physical and psychological limitations.

(Reproduction: “Robot: Mere Machine to Transcendent Mind”, by Hans Moravec)

The comparison used by Moravec has two indicators: the first, the megabyte, as a data storage unit, and the second, the MIPS (Millions of Instructions per Second), which is a highly questioned indicator of computational performance capacity. In a graph plotted with the megabytes indicator horizontally and MIPS vertically, the human brain is in the upper right corner of the graph, with a high capacity for data storage and a high condition of processing instructions per second.

The comparison is often accompanied by a graph of computational processing power cost. This graph shows that for $ 1,000 we have been able to acquire machines with increasing processing power over the years, following Moore's Law [which said, in 1965, that the processing power of computers would double every 18 months]. It is from the simple and superficial reading of this graph that comes most of the predictions that from 2030 to 2050 we will have machines with processing capacity similar to that of the human brain.

Reproduction: “Robot: Mere Machine to Transcendent Mind”, by Hans Moravec)

Some important points of reflection little done in reading this information. This instruction processing capability does not take into account the complexity of the instruction given, nor the integration of the various instructions necessary to handle the various complex tasks that, as humans, we perform every day.

I don't mean that in 2030 we won't have computers, machines or robots capable of doing much of the complex tasks that we humans do, nor do I mean the opposite. I just think that the comparison and the indicators used are too simplistic for us to reach any credible conclusion.

The key point is that the discussion is much broader and more complex than that, and begins with a very simple and profound question: "What is intelligence?"

Taking the entry in the dictionary available on Google, we have:

intelligence:

feminine noun

1.

faculty to know, understand and learn.

2.

ability to understand and solve new problems and conflicts and to adapt to new situations.

3.

set of psychic and psychophysiological functions that contribute to knowledge, to the understanding of the nature of things and the meaning of facts. "the disease affected his intelligence"

4.

way of interpreting, of judging; interpretation, judgment.

5.

BY METONY

highly intelligent individual; sumity.

6.

harmony, mutual understanding. "live in good intelligence"

7.

secret agreement or combination; machining, collusion.

Focusing on the first four definitions, which are the most common and comprehensive, I dare to say that we already have in hand a good part of the necessary prerequisites to create algorithms, machines and robots to which we can definitely give the nickname of intelligent machines. And if we continue at this pace, it is likely that we will have machines and computers capable of integrating an ever-expanding range of instructions and activities, being able to do increasingly complex activities.

The technologies that today are behind what we call in the popular jargon of artificial intelligence allow us to do a set of types of tasks that combined can be the basis of truly intelligent artificial beings.

If human intelligence today allows us to:

Understand context, Extract relevant information from a conversation, Plan and optimize, Speak, Interact, Respond appropriately to a problem, Generate phrases and stories, Recognize people and objects, Negotiate, Learn patterns, Follow established rules

The technologies below, all linked to what we call artificial intelligence, allow us to do the same things respectively:

Computer vision and semantic visual segmentation, Natural language processing, Planning and optimization algorithms, Artificial voice generation, Chatbots, Expert systems, Natural language generation, Facial and object recognition, Optimization and decision-making algorithms, Learning machine and deep learning, Robotic Process Automation - RPA

And at this point there is a big disagreement with me regarding Hans Moravec's chart. What will make artificial intelligence algorithms more and more “intelligent” and capable of handling the highly complex tasks that we humans can perform is not the processing capacity, not even the storage capacity, but rather, our ability, as humans, to understand the full potential of this technology so disruptive that it is artificial intelligence, and to use our creativity to build increasingly innovative architectures, systems and solutions.

The technological bases for the supercomputer or super-robot are already present today. It is up to us to organize these technologies, not to do what humans are able to do, but to do things we never dreamed of.

If you want to help build this new world that we call the future, it is definitely time to put a lot more knowledge about this technology in your daily life. And may the future come!