Nowadays, it’s as if every article about digitalization is casually peppered with the terms ‘artificial intelligence’ and ‘machine learning’. They are almost always used together and without any extra explanation. I’ve recently spent some time reading up on the subject to finally understand how the two differ.
The term artificial intelligence (AI) originates from computer science and refers to the capacity of machines to perform human-like tasks on their own. AI covers all kinds of computer intelligence such as speech recognition, computer vision, virtual assistance, graph analysis, robotic process automation, chatbots and, last but not least, machine learning.
In other words, machine learning is a form of AI. It uses algorithms and statistical models to recognize patterns and perform tasks without explicit instructions. Software equipped with machine learning can recognize how the weather affects customer demand or how the traffic situation affects deliveries, for example. Companies can then take such effects into account in their forecasting and planning.
Endless AI discussion
I sometimes wonder why authors mention AI and machine learning in the same sentence and without any explanation. Do they assume that the readers – like themselves – don’t know the difference anyhow? It remains absolutely imperative to clarify the meaning and implications of both terms. Otherwise we will quickly get caught up in the endless discussion of how robots and computer systems are taking over all human processes in factories, warehouses and offices.
The managing director of an international manufacturing company recently asked me how his company could utilize machine learning. I replied by asking him what the computer needed to learn: which huge volumes of data is his company generating that a self-learning computer could help to make sense of?
The supply chain director of a global restaurant chain that also sells its branded sauces through supermarkets told me that his department used to have to do a lot of repetitive, time-consuming data input, but has since automated much of the human keyboard action using robotic process automation (RPA).
In other words, AI is an extremely promising toolbox offering countless variations, but supply chain professionals need to maintain a clear focus on which problem they want it to solve.