The objective of this project is to develop an algorithmic system, trained on both public legal data and proprietary negotiation data, that will provide tailored predictions of likely court-based dispute resolution outcomes and optimal settlements. In the proposed sub-project, a team of interns (JD students) will be modelling and labelling dispute-resolution data on complaints against hospitals and municipalities. We anticipate that a first dataset, ready for algorithmic model testing, will be completed by the end of the internship.