PREFACE

This publication presents the proceedings of the 20th International Trust Workshop, which was
co-located with AAMAS/IJCAI/ECAI/ICML 2018 and held in Stockholm, Sweden on July 14
2018.

Trust is important in many kinds of interactions, including computer-mediated human
interaction, human-computer interaction and among social agents; it characterizes those
elements that are essential in social reliability. It also informs the selection of partners for
successful multiagent coordination. Trust is more than communication that is robust against
repudiation or interference. Increasingly, there is concern for human users to trust the AI
systems which have been designed to act on their behalf. This trust can be engendered through
effective transparency and lack of bias, as well as through successful attention to user needs.
Mistrust has also emerged as a current theme, especially within online settings where
misinformation may abound. AI approaches to addressing this concern have thus come into
focus.

This workshop aims to bring together researchers working on related issues regarding trust and
artificial intelligence, expanding the discussion beyond the borders of multiagent trust
modeling, where research and dialogue has been very active over the past twenty years.

Many computational and theoretical models and approaches to reputation have been developed
recently. Further, identity and associated trustworthiness must be ascertained for reliable
interactions or transactions. Trust is foundational for the notion of agency and for its defining
relation of acting "on behalf of". It is also critical for modeling and supporting groups and
teams, for both organization and coordination, with the related trade-off between individual
utility and collective interest. The electronic medium seems to weaken the usual bonds of social
control and the disposition to mislead grows stronger; this is yet another context where trust
modeling is critical.

This workshop continues the tradition of bringing together at AAMAS each year researchers
working on modeling trust in multiagent systems but with an expanded vision for the gathering,
to encourage participation from researchers working on related issues: regarding trust of AI
systems (and the need to address possible mistrust of these systems) and regarding concern
about mistrust for applications where AI solutions may be of use (such as the web and online
social networks).


Workshop Co-Chairs

Robin Cohen - University of Waterloo, Canada
Murat Sensoy - Ozyegin University, Turkey
Timothy J. Norman - University of Southampton, UK