Artificial intelligence

Artificial intelligence (AI) itself is an umbrella term for “smart” technologies that are aware of and can learn from their environments in order to assist and/or augment human decision making. Subfields include natural language processing, machine learning, deep learning, vision and pattern recognition, speech recognition and many others.

The most distinguishing aspect that separates AI from other general-purpose software is that it can enable machines to respond on their own to signals from the external world–signals that programmers do not directly control and therefore cannot anticipate.

Features

Uses tools and insights from many fields, including computer science, psychology, philosophy, neuroscience, cognitive science, linguistics, operations research, economics, control theory, probability, optimisation, and logic
Fields include natural language processing, machine learning, deep learning, vision and pattern recognition and many others
Applied cross-industry to solve problems or make processes more efficient

AI in practice

AI can find sales leads by emailing in natural language on your behalf and hand off the interaction after the converting them to customers. Using insights from AI, companies are able to better forecast sales by customer and as a result better target the higher value customers.

In 2016, a global technology company tested an experimental self-driving car with a neural network that learned entirely by watching footage of human drivers - for example recognising the outlines of roads without any explicit instruction from researchers. More conventional uses of AI in autonomous vehicles include vision and pattern recognition to process and act upon data captured by sensors attached to a vehicle.

An international entertainment company uses machine learning to process a wide and deep body of data about both viewers and content, ultimately automatically recommending titles based on a user’s past viewing behavior and similarity of content to other content. There are two primary business benefits: making better use of the content library and obscure “long tail” titles, and strongly improving viewer engagement.

A worldwide online payments processor, uses AI pattern recognition technology to look for unusual behavior that may signal fraud. It examines both a user’s own transaction habits as well as patterns of likely fraud stored in its database; comparing each transaction against thousands of potential problem scenarios. As a result, the company has a fraud rate of 0.32% of revenue vs. the 1.32% average of other merchants.