In recent years interface agents have been developed to assist users with various tasks. Some systems employ machine learning techniques to allow the agent to adapt to the user's changing requirements. With the increase in the volume of data on the Internet agents have emerged that are able to monitor and learn from their users to identify topics of interest. One such agent described here has been developed to filter mail messages. We examine the issues involved in constructing an autonomous interface agent that employs a learning component and explore the use of two different learning techniques in this context.

This paper investigates factors influencing public transport passengers’ pre-travel information-seeking behaviours in a British urban environment. Public transport traveller surveys were conducted to better understand the journey stages at which information was sought and the information sources used. A multivariate explanatory model of pre-travel information-seeking behaviour was developed using binomial logistic regression. Explanatory factors considered include socio-demographics, trip context, frequency of public transport use, information sources used, and smartphone ownership and use. Findings suggest that travel behaviour (5 + trips weekly, and < 1 trip weekly), socio-demographics (unemployment/unknown employment), trip context (journey planning stages, mode of transport), and preferred information sources (Internet site, word-of-mouth, visits to travel shop/centre/library) were significant predictors of pre-travel information-seeking behaviours among surveyed travellers. While the final model found that bus users are significantly associated with the use of Internet sites as a source of pre-travel information, rail users rely significantly on a multiplicity of sources comprising Internet sites, word-of-mouth, and visits to a travel shop/centre/library. The final model suggests that metro (light rail) users tend not to seek pre-travel information. The odds of seeking pre-travel public transport information are 2.512 times greater for respondents who reported < 1 trip per week as opposed to those who reported 5 + trips per week. These findings are relevant for passenger information strategies deployed by operators and authorities and can be used to caution against a “one size fits all” strategy for travel information service provision. Implications and suggestions for future research are discussed.

Internet of Things (IoT) devices such as connected sensors are increasingly being used in the public sector, often deployed and collecting data in public spaces. A theme commonly seen in the rhetoric surrounding public space IoT initiatives is empowerment, and these deployments are broadly perceived as beneficial by policy makers. However, such technology presents new governance challenges. It is important to ask who is empowered and who benefits, and we must ensure that such technological interventions follow democratic principles and are trusted by citizens. In this paper, we investigate how risk, transparency, and data governance require careful consideration in this domain, describing work which investigates how these combine to form components of trusted IoT ecosystems. This includes an overview of the landscape of public space IoT deployments, consideration of how they may often be subsumed in idealized smart city focused rhetoric, and discussion of how methodologies such as design fiction in community settings can uncover potential risks and concerns. Our findings suggest that agency, value and intent associated with IoT systems are key components that must be made transparent, particularly when multiple actors and stakeholders are involved. We suggest that good governance requires consideration of these systems in their entirety, throughout the full planning, implementation, and evaluation process, and in consultation with multiple stakeholders who are impacted, including the public. To achieve this effectively, we argue for transparency at the device and system level, which may require legislative change.