Using the development of a computer system by a collaboration between the Technical University of Denmark and IBM, called DTU-Watson, this paper investigates the usage of machine learning in contract drafting and negotiations of contracts. Using the system as a case study, we discuss a number of implications related to the application of machine learning technology to legal work: I) whether machine learning will reduce the need for lawyers in contract negotiations; II) the difficulties encountered by lawyers in exploring the potential of new technologies; and III) broader consequences in terms of organization of legal offices and ownership of technology.