Ashwin Vaidya About
Originally published on 10th September 2018|Last Updated: 01 December 2023

What is Artificial Intelligence

Table of Contents

Photo by Franck V. on Unsplash

What is Artificial Intelligence?

Artificial Intelligence! The solution to all our problems or our future overlords? Whichever side you are on, this article attempts to broadly explain what comprises artificial intelligence.

There are a lot of definitions for what artificial intelligence is but, for this discussion, I will define it as the study of intelligent machines.

If you are a film/TV buff then AI may bring to mind images of R2D2 from Star Wars or HAL 9000 from 2001 Space Odyssey. However, AI need not be limited to robots with human-like intelligence. It can also be a software which performs a particular task; as seen on our smartphones, from maps which show us the shortest routes; or shopping/video apps which suggest products/movies based on our interests.

So, How Does One Define AI?

Formally, it can be explained as the study of intelligent agents which perceive their environment and then take actions which maximize their chance of achieving their goals.

Consider the example of an autonomous car. Its goal is to drive from one location to another, while avoiding all obstructions in its route. Here, the car is an intelligent agent and its environment comprises the road and obstacles.

What characterises An Intelligent Agent?

An intelligent agent uses sensors to determine the state of its environment. In the example of the autonomous car, cameras, RADARs, LiDARs etc, are the sensors. The agent then used its knowledge to decide the response and takes action using actuators (motors in case of the car) to interact with the environment. That is, move forward, apply breaks, take a left, etc.

Where Does The Knowledge of the AI Agent Come From?

The most elementary AI systems have a knowledge base that comprises of explict, hard coded input-response pair rules. Such systems are also called as Expert Systems. Now, AI is transitioning towards an implicit representation of knowledge. These AI agents do not have input-response pair hardcoded but rather they figure out the response based on the data acquired by them during training. This is where Machine Learning and Deep Learning come in. I will explain both these terms in a different post.

If you want to read more on AI environments refer to this post.

Before I end this post, let me quickly explain a few more terms.

Artificial Narrow Intelligence (ANI) – This is also referred to as weak intelligence. This type of AI specializes in only one area. Route finding in Google Maps is an example of ANI. In fact, most of the AI we are surrounded by today fall into this category.

Artificial General Intelligence (AGI) – This is also referred to as strong intelligence. This type of intelligence is human-level intelligence and is the goal of AI researchers. Such an AI is good at a diverse set of tasks, can comprehend complex ideas and learn from experience.

Artificial Superintelligence (ASI) – This type of intelligence is far superior to that of humans.

You have reached the end 🙂