A Concise Definition of AI Terms

I first encountered the term “AI” in the mid-1990s while studying computing. At the time, the fundamental concepts were similar to what they are today, but due to limited computing power, achieving them was more of an aspiration. Beyond simple “AI,” innovations like ChatGPT were entirely out of reach. Nowadays, the term “AI” is used much like how we refer to “Cloud” or “IoT” — as a broad, encompassing buzzword.
I’ve attended events like NRF (the Retail Show in New York), where “AI” was mentioned in nearly every talk I attended. However, what many speakers referred to as “AI” often didn’t align with its true definition. I’ve also been part of numerous conversations where people have asked, “Can we label this as AI?”—primarily to include the buzzword in their product’s title.
To bring clarity, I sat down with a Data Science colleague, and together, we identified three key definitions of AI:
AI
Artificial Intelligence (AI) can be defined in many ways. It is an umbrella term for a range of algorithm-based technologies that solve complex tasks by carrying out functions that previously required human thinking. Decisions made using AI are either fully automated or with a ‘human in the loop’.
Machine Learning
Machine learning is the analysis and construction of algorithms (models) that can learn from and make predictions on data. The more the model is trained on (relevant) data, the more accurate it will get in performing its task, for example playing checkers.
Deep Learning
This subset of machine learning uses a neural network with multiple layers. Neural networks originated from attempts to simulate the behaviour of the human brain. Such models do not (yet!) match the ability of the human brain (artificial general intelligence, another name for super intelligence)), but can “learn” from large amounts of data.
When most people talk about AI, they are quite often thinking it is Deep Learning, but what they actually have is more like the general term AI.