Learning to love the bots

Artificial intelligence (AI) has spread its wings: from Tesla's self-navigating cars, to Google's DeepMind defeating the world champion at Go (a board game more complicated than chess), to IBM's Watson diagnosing diseases. Software is moving beyond executing black-and-white instructions to making complex and often subjective decisions. In 2017 society will begin to trust machines in much the same way that we trust people. AI introduces a kind of unpredictability not found in traditional software. In solving complex tasks, it is too difficult to build rules to cover every eventuality. Instead, AI systems learn from experience.