AI, Machine Learning, and Deep Learning in Plain English

Artificial Intelligence (AI) is the older of the three terms. Machine Learning (ML) came later. Deep Learning (DL) is the most recent term.

Everything related to ML is a part of AI, but not the other way around. Everything related to DL is a part of ML, but not the other way around.

AI is about artificial systems that mimic natural intelligence in some sense. ML is a technique used for building AI-based systems. DL is an ML technique.

ML techniques use the input and output data of a system to learn to mimic that system in some sense.

Neural Nets (NN) are an ML technique. They are layered structures. Their first layer is where the inputs are applied. The final layer is where the computed outputs are made available. The layers that reside between the input and output layers are called the hidden (neuron) layers as they are not directly connected to the outside world.

NNs having a single or no hidden layer are called Shallow NNs. NNs having two or more hidden layers are called Deep NNs. DL relates to techniques used for training Deep NNs.

NOTE: AI and ML are sometimes used interchangeably, and there is nothing patently wrong with that! We should just try to deduce the intended implication from the context.