Understanding the basics of artificial intelligence

Objective of this article

Many of you reading this cannot escape to have noticed the exponential rise in businesses across almost all sectors claiming to now have “AI-enabled” solutions.? Indeed, many have jumped on this growth not because they actually have such solutions or even understand artificial intelligence but simply do not want to be seen to be lagging behind.

Against this backdrop, businesses are making choices to buy services from such companies offering AI-enabled solutions but do these buyers actually understand artificial intelligence?? Are they in a position to make a considered evaluation of what is being offered?

This article is an initial attempt to help readers understand what artificial intelligence is, its history, where it is now and where it is heading.? I stress that this is only an introduction.? Further articles will explore the technicalities in more detail.? Let’s begin with trying to define human intelligence.

What is human intelligence?

Human intelligence is the intellectual power of humans, which is marked by complex?cognitive?feats and high levels of?motivation?and?self-awareness.

Intelligence enables humans to remember descriptions of things and use those descriptions in future behaviours. It is a cognitive process.

It gives humans the cognitive abilities to?learn,?form concepts,?understand, and?reason, including the capacities to?recognize patterns, innovate,?plan,?solve problems and employ?language?to?communicate. Intelligence enables humans to?experience?and?think.

Learning refers to the act of retaining facts and information or abilities and being able to recall them for future use, while intelligence is the cognitive ability to perform these and other processes.

What is artificial intelligence?

Put very simply, artificial intelligence is the ability of a computer or computer-controlled robot to perform tasks commonly associated with intelligent beings.

The image below depicts aspects of intelligence.? AI is not capable of all of these.? The ones in lighter green shading are quite possible and currently in use.? However, the ones in darker green shading are a very long way off and it is largely highly unlikely that a machine can be created to apply these abilities by itself or of its own volition

There are three basic types of AI:

1. Cognitive simulation where computers are used to test theories about how the human mind works in processes such as memory recall or facial recognition.? This is currently used in neuroscience and cognitive psychology.
2. Advanced information processing, also known as Applied AI. It aims to produce commercially viable systems such as in medical diagnosis and loading ships.
3. Artificial general intelligence, also known as Strong AI. It aims to build a machine whose overall intellectual ability cannot be distinguished from human ability.? This is what the hype is all about.? But currently, we cannot even replicate the intelligence of an ant or a cockroach.

The focus of AI research

Following my earlier assertion that AI is not capable of acquiring all attributes of human intelligence, AI researchers and developers have focused on just 5 areas of intelligence:

1. Learning
2. Reasoning
3. Problem Solving
4. Perception
5. Language

Here are some basic descriptions

Rote learning: This is learning by trial and error such as in chess playing programs.? The computer selects moves at random until it finds one that succeeds.? It then stores the position of the pieces and the solution.? If the same position is encountered in future, it recalls and applies the solution

Generalisation learning:? This applies past experience to similar, new situations such as changing verb tenses.? The computer is programmed to add “-ed” or “-ing” to the present tense of a verb based on earlier programming of a different verb root.? For example, it has been taught that the past tense of the verb “walk” is “walked” and the present participle is “walking”.? It then knows that the past tense of the verb “talk” is “talked”.? However, this is very difficult to apply to irregular verbs and across different non-Latin based languages.

Deductive reasoning:? Apples are either red or green.? If the apple is not red, it must be green. The truth of the statements guarantees the truth of the conclusion.? However, what happens if a new apple is developed which has a yellow skin?

Inductive reasoning:? ?Previous car accidents like this were caused by driver error.? Therefore, this accident was caused by driver error.? Here, the truth of the statement only supports the conclusion.? New evidence can alter the reasoning within the model and its subsequent conclusions

Special purpose problem solving:? This is designed to solve a particular problem such as which washroom should I use if I am female as opposed to male?

General purpose problem solving: Programs can solve problems involving multiple options to achieve the desired outcome.? In route finding GPS will use starting and intended destination for any desired routes. The program then uses simple instructions such as “turn right”, “turn left”, “continue ahead”, distance etc.

Perception:? Artificial sensory “organs” scan the environment around them. The computer then decomposes the “scene” into various spatial relationships.? It then “digitalises” the image. Combined with a robot, the computer can then reconstruct elements into a physical whole.

Language:? Language is a system of signs having meaning and not confined to speech.? Human languages can formulate an unlimited variety of messages.? Large language models such as ChatGPT do not understand language.? They can respond only as programmed based upon best probability of the required response.

Future articles

So that’s all for now.? To keep these articles short enough to be readable I am separating the topics across a number of articles.? The forthcoming ones will address:

• The origins of AI going back to and beyond 1935
• Symbolic vs connectionist approaches to AI
• Evolutionary computing, microworlds and expert systems
• Artificial neural networks
• Modern AI – machine learning and autonomous vehicles; virtual assistants and large language models
• Is Artificial General Intelligence (AGI or Strong AI) possible?
• What does all this mean for AI in the health financing ecosystem?

Also, look out for my future articles on quantum computing and how this could accelerate AI beyond our imaginations!