For 70 years we have been arguing about whether computers will ever equal us in intelligence. Some gurus fear the ‘Singularity’, when computers will be so intelligent that they will enslave us – if we are lucky. But a greater danger is that we will enslave ourselves to stupid computers: the ‘Surrender’. The recent startling successes in machine intelligence using a technique called ‘deep learning’ have given new urgency to this tension. Deep learning has put computers in touch with human culture as never before with dramatic results. Nevertheless, it is still impossible to foresee a time when machines will be sufficiently embedded in society to be independent of human input or when we cannot distinguish between the social and linguistic understanding of humans and computers. Fictional portrayals of intelligent machines simply confuse the issue. Anyone can try the simple elements of a Turing test provided here: you can show that artificial intelligence is not as clever as it is said to be. | Artifictional Intelligence by Harry Collins, Paperback | Indigo Chapters