In this paper we describe our experience conducting an ‘ethical hackathon’ to promote the ethical design of AI systems. The model of the ethical hackathon has been developed by researchers in the Human Centred Computing theme as a novel twist on the conventional hackathon competition. Ethical hackathons are fun, educational events in which interdisciplinary teams compete on a design challenge that requires them to consider how responsibility mechanisms can be embedded into what they are building.

The ethical hackathon described in this paper was part of the UnBias project. In the paper we highlight the potential for these events to foster the ethical design and development of AI systems but also identify some practical challenges in running them. We conclude that a successful ethical hackathon needs to foster genuine interdisciplinarity and carefully manage participant expectations. We build on our own experiences by suggesting ways to optimise the ethical hackathon model.