Industrialisation of AI offers great promise and some peril

Private companies, not universities, are pioneering artificial intelligence research and reaping almost all the gains

On the peanut butter spectrum, artificial intelligence is rapidly moving from smooth to crunchy. Many of the technology’s most high-profile uses to date have been in the fields of games playing, image recognition and language generation, which we can all recognise. AlphaGo, developed by Google DeepMind, beats the strongest players in the ancient game of Go. Generative adversarial networks create deepfake videos substituting Burt Reynolds’ face for that of Sean Connery in a James Bond film. Open AI’s GPT-3 program can write an unnervingly authentic poem in the style of Emily Dickinson. But AI is increasingly being used in many more specific, invisible and productive ways across industry and in the hard sciences. Perhaps more narrowly defined in most cases as machine learning, AI has become a powerful technological tool that is being used to optimise search engines, accelerate drug discovery, invent new materials, improve weather prediction and deepen our understanding of mathematics, biology, chemistry and physics. That diffusion of AI into almost every corner of the economy and society is going to profoundly affect us all.

Read more