The problem of 'truth discovery' is to find the true facts concerning a number of objects when faced with conflicting reports from data sources of unknown trustworthiness and reliability. This is particularly important in today's 'post-truth' period, where conflicting information can be found in abundance on the web and social media platforms.

Many different algorithms for truth discovery have been proposed in recent years. My work has instead involved looking at the problem from a theoretical point of view, and applying the axiomatic method of Social Choice to evaluate algorithms with respect to their theoretical properties.

In this talk I will introduce the problem and go over my initial work in this area.