Universities are digital machines these days. But many of the decisions that have to be made as a result are not technical at all. They are about the nature of research and its public benefits, about how learning and teaching takes place, and how we confront difficult ethical issues. Strategic choices that are made now will have significant implications for the ways in which knowledge will be created and shared in the future. Paywalls and passwords are an irritating but natural consequence of the privatisation of the web; they are also essential for online security. For the future of research, though, the need for openness is far more than a convenience. It arises because the volume and rate of production of online publications and digital data sets has now outgrown the limits of conventional research methods and is changing the ways in which new knowledge is created. Without openness across global digital networks, it is doubtful that large and complex problems in areas such as economics, climate change and health can be solved. Appropriate and effective policies and protocols that ensure openness are essential in ensuring access to both research publications and massive digital data sets. Automated searches of the publication record are not feasible if subscription requirements and paywalls constantly prevent access to the copy-of-record. This is why the current debate about "green" open access – delivered by repositories – and "gold" open access – delivered by journals – and article-processing charges is so important. Our universities were built from the principles of scarcity and closure: restricted access to libraries; special knowledge that could only be passed on in the lecture theatre; closed communities of scholarship. Today we are drowning in digital information, available almost everywhere via interfaces and cheap devices that know where we are. Learning and teaching is about making sense of this bewildering open world, with all its opportunities and its increasing dangers.