Four Approaches to AI

 Four Approaches to Defining Artificial Intelligence

Assoc Prof Sonika Tyagi
Data Science & AI | RMIT School of Computational Technology| Australia

When we ask “What is AI?”, it turns out there isn’t just one answer. In fact, there are four broad ways to define and understand AI: based on whether we want machines to think or act, and whether we want them to do so like humans or like ideal rational agents.

Let’s explore each of these four perspectives.

1. Acting Humanly: The Turing Test Approach

One of the earliest and most well-known definitions of AI comes from Alan Turing. In 1950, he proposed what we now call the Turing Test, an operational test of a machine’s intelligence.

Turing suggested that if a machine could carry out a conversation well enough to convince a human that it too was human, then we could say the machine was intelligent.

In this view, intelligent behavior is about replicating human-like performance, esp in language, reasoning, and adaptability.

Now, when we say “acting humanly,” we don’t mean humans are perfect. In fact, quite the opposite... humans make mistakes, take shortcuts, and use heuristics. The point is not to build a flawless machine, but to build one that behaves in a plausibly human way.

Discussion prompt: Can you think of an example of an AI system that acts humanly but does not think humanly?

2. Thinking Humanly: The Cognitive Modelling Approach

This approach goes deeper. Rather than just copying behavior, it asks: Can we build systems that think the way humans think?

To do this, we need to understand how the human mind works. This is where AI intersects with cognitive science, neuroscience, and psychology. There are three main ways to study human thought:

  1. Introspection:- examining our own thought processes

  2. Psychological experiments:- observing how people solve problems

  3. Brain imaging:- studying the biological basis of cognition

Once we have a theory of how humans think, we can build a computer program that implements it. If the program behaves like a human, not just in output, but in timing and reasoning patterns then we say it’s an example of cognitive modelling.

              

 Discussion prompt: What are the limitations of Thinking Humanly approach?

3. Thinking Rationally: The Laws of Thought Approach

This approach goes all the way back to ancient Greek philosophy. Aristotle tried to formalize the principles of correct reasoning, known as logic.

A classic example of this is a syllogism:

Socrates is a man. All men are mortal. Therefore, Socrates is mortal.

In this tradition, AI should be about systems that think logically and draw correct inferences. If we know what is true, and we follow valid reasoning steps, we should arrive at true conclusions.

While elegant, this approach has practical challenges:

  • It assumes perfect, complete knowledge, which rarely exists in real-world scenarios

  • It can require too many computations to be feasible in real-time

  • Most importantly, it focuses only on thinking, not on acting

And that brings us to the final approach…

4. Acting Rationally: The Rational Agent Approach

This is the most modern and widely adopted definition of AI today.

An agent is any entity that perceives its environment and acts upon it to achieve its goals. An AI system is rational if it acts in a way that maximizes its chances of success, based on what it knows.

For example, imagine a thermostat agent. Its goal is to maintain a room at a desired temperature. It uses sensors to perceive the current temperature and then takes action i.e. turning the heater or cooler on or off to reach its goal.

Agents can take many forms:

  • Human agents:- like us

  • Robotic agents:- like autonomous drones or self-driving cars

  • Software agents:- like recommendation systems or virtual assistants

In contrast to the “laws of thought” approach, which emphasizes correct inference, the rational agent approach focuses on outcomes. Sometimes, acting quickly or using heuristics leads to better results than perfectly logical reasoning.

For instance, pulling your hand away from a hot stove isn’t something you calculate...it’s a reflex. But it’s still rational.

Final Thought:

AI systems don’t always need to mimic human thought or behavior to be intelligent. Sometimes, they just need to make good decisions under uncertainty. And depending on the application, one approach may be more suitable than another.

Understanding these four perspectives i.e. acting humanly, thinking humanly, thinking rationally, and acting rationally, gives us a framework to appreciate the richness and diversity of AI.

Comments

Post a Comment

Popular posts from this blog

Modelling genomic language using NLP and LLMs

Image
Author : Navya Tyagi  Genomic data consists of DNA, RNA, and protein sequences that can be represented as strings of unstructured text. These sequences can be very large in size. For instance, human DNA is made of 3 billion A,G,C, and T letters. There are hidden patterns that can be considered equivalent to "words" in a natural language. But all of these words are not known and more importantly the grammar that genomic lanaguege follow is not well understood. These biological words with critical functions are of interest to study disease and development processes. Sometime a mutation in these "words" may result in a disease condition. Determining these "words" with biological function is a computational challenge. Figure 1: a) Code of DNA can be written using letters A,C,G,T. A pairs wiht T and G pairs with C making a double stnraded (helical) structure out of two DNA strands. b) message from DNA is transcribed to RNA (also known as messenger RNA). The ...
Read more

Machine Learning: Interpretability vs Explainability

Image
 Authors: Navya Tyagi & Esha Srivastava, Tyagi Lab, RMIT University, Melbourne, Australia   Introduction   We are aware that machine learning models are built upon vast amounts of pre-existing context and are constantly learning. The advent of machine learning is meant to assist humans in making complex decisions. However, there are still many misconceptions and uncertainties about how much trust one should place in machine learning when it comes to decision-making—especially when these decisions could impact an individual’s life. But what if there was a way to at least understand or backtrack how machine learning makes decisions? Hence, in this blog, we are going to discuss interpretability and explainability. We often come across these terms and may use them interchangeably. However, it is important to understand the distinction between them. Interpretability refers to converting the implicit information within a neural network into human-interpretable insights. Th...
Read more