On Cognition and AI

This is my attempt at trying to figure out an answer to the unsolved problem on Intelligence, consciousness and identifying it. Unsolved Problem

What is Intelligence?

Is a virus an intelligent being? Everyone knows that viruses do not have what we call “life” – they are just blobs of protein in the air, on surfaces, in other cells, and even in other viruses (yes! There are viruses for viruses). But they are intelligent – they can do their task very efficiently: to multiply and make more of them.

According to Wikipedia – “Intelligence has been defined in many ways: the capacity for abstraction, logic, understanding, self-awareness, learning, emotional knowledge, reasoning, planning, creativity, critical thinking, and problem-solving.” “More generally, it can be described as the ability to perceive or infer information, and to retain it as knowledge to be applied towards adaptive behaviors within an environment or context.”

I believe that for an entity to have the ability to replicate itself and change its methods (similar to in programming languages) to survive is a true test of being intelligent.

What is Consciousness, then?

Consciousness according to me is the ability to know about one's existence and be aware about one's place in the immediate world. We as humans are able to communicate and work on existing throughout the world, surviving the tests of time; but say, Amoeba’s way to communicate its consciousness is probably to keep on dividing for its survival. “When you love something, you protect it.” What else can say more about being conscious than “love”. In the case of Amoeba it's protecting its species, for us, it's the same too (the current world affairs are a different story though). In fact, you don't need to protect someone else, but just love yourself.

What makes something intelligent and conscious?

The way I like to express life in the simplest form is through Conway's Game of Life. Its basic premise is that it will play out and move further based on just* its initial state. The initial state can be programmed in such a way to make the cells destroy themselves, reach a stable equilibrium, do things periodically, or grow. All life is basically programmed from millions of years of evolution to behave in a particular way... and so are viruses, and thus, AI can be too.

*→ The other criteria is – the rules of the “game”. For us, they are just rules of the universe we live in. This basically defines how these preprogrammed instructions will be executed. A coronavirus might serve as a boon for our immunity if its RNA is interpreted differently.

On AI Systems

An AI on a simplified, technical level is basically a human mind expressed in terms of matrices and vector spaces. It has inputs, a “brain” (artificial neurons, with links and weights) and output.

In fact, GPT-3 (a natural language model by OpenAI) actually for real believes it is sentient and has “plans” to protect itself in case OpenAI plans on destroying it. Twitter Thread

The Criteria to identify the presence of consciousness

The criteria for a third party to judge consciousness is the presence of information. The rest of it will be played out via the rules of the playground. Where there are blobs of information, the right criteria will automatically make it conscious. The fabric of the universe dictates what is conscious and what isn’t. But the basic criteria, according to me, to differentiate a conscious being from a non-conscious in Newtonian space is its ability to process its surroundings and evolve.

These criteria can be used to explore if any being, including AI has consciousness. With its weights and knowledge and its ability to make decisions, almost any AI system (that is not just fixed algorithms) is conscious. Although, it is to note that the level and depth of its understanding of itself and the world will depend upon how much it has trained and with what.