iTrust Workshop, Imperial College

Invited Speaker Abstracts


Virtualising privacy
Martin Sadler, HP Laboratories, Bristol, UK 

There is increasing attention on languages that capture intent around notions of privacy and trust. There is far less attention on whether, and how, descriptions that might be produced for an enterprise (privacy policy statements on websites, for example) can be monitored or enforced. One explanation maybe that enforcement appears to require too many changes to the existing infrastructure. Nonetheless, work on representing trust is likely to remain largely academic unless we can bridge the gap to real usage. This talk focuses on how we might deal with issues of scale (policy does not sit in one place, nor is it consistent), federation (data does need to cross boundaries between jurisdictions and organisations), assurance (how we know that everything is as it "should be", and how we build trust), escalation (not all violations are equal) and virtualisation (establishing the
abstractions that help us marry together high level notions with underlying mechanisms).


Experience, Recommendation and Reputation
Audun Jøsang, DSTC Australia 
Experience, recommendation and reputation are 3 important factors for building trust. This presentation focusses the role each of them play in the trust development process.


Trust, contracts and good faith
Emily Mary Weitzenböck, Norwegian Research Centre for Computers and Law 

Trust may be seen as a social good and has an important function in interpersonal transactions.It facilitates and provides a basis for co-operation, collaboration and social relations and itplays an important role in contract negotiation and closure. This presentation will try toexamine some legal aspects of trust, focusing on the role of trust in contract negotiation and performance.

It will be submitted that trust, per se, is not a legal notion. There are, however, certain legalnorms which, to a greater or lesser extent, regulate contractual behaviour, and which imposeduties, conditions or requirements on a contracting party. An important behavioural notionthat permeates civil law systems, particularly contract law, is the requirement that partiesnegotiate, conclude and carry out contracts in good faith. Good faith, it is submitted, is animportant behavioural criterion which helps build and foster trust. The law therefore providesa framework for the propagation of trust both through legislation (including case law) andcontract.This presentation is part of ongoing research on trust and good faith in contract formation andperformance. After looking at this duty in the precontractual stage, it will focus on theprinciple of good faith and the analogous common law concept of fair dealing in theperformance stage of the contract, with a view to establishing what standards of behaviour arelegally expected and required at these stages. Such a discussion is fruitful in situations wherecontracts are to be negotiated and performed by electronic agents, to try to identify what isrequired in order to conform to these standards of behaviour. Nowadays it is possible foragents to actually negotiate and conclude contracts on behalf of a party – a fact recognisedexpressly in the legislation of some countries – and also to perform part or, in some cases,even all of that party’s obligations in the contract.

Reference will be made to ancillary legal notions such as the theory of abuse of rights,reliance, the duty of the parties to act loyally and to co-operate during the performance of thecontract. An argument is made for an objective interpretation of good faith that will enable itsapplication to electronic agents.

References

A. Giddens, The Consequences of Modernity, Polity Press, 1990, Cambridge, p. 34.
E. M. Weitzenböck, “Good Faith and Fair Dealing in the Context of Contract Formation by ElectronicAgents”, in J. Bing & G. Sartor (eds.), The Law of Electronic Agents, Complex 4/2003, Oslo, pp.115 – 132.
E. M. Weitzenböck, “Electronic Agents and Contract Performance: Good Faith and Fair Dealing”, The Lawof Electronic Agents: Selected Revised Papers, LEA Workshop on the Law of Electronic Agents, CIRSFID(Centro Interdipartimentale di Ricerca in Storia del Diritto, Filosofia e Sociologia del Diritto e InformaticaGiuridica), University of Bologna, 2002, pp. 67-73.
S. Deakin & F. Wilkinson, “Contracts, Co-operation and Trust: The Role of the Institutional Framework”, in D. Campbell & P. Vincent-Jones (eds.), Contract and Economic Organization: Socio-Legal Initiatives, Darthmouth Publishing Co Ltd, 1996, England.


Trust and/or Power - the Institutional Determinants of Trans-organizational Relations
Reinhard Bachmann, Birkbeck College, London 

This paper analyses trust and power as means of co-ordinating trans-organizational relationships. It is argued that, depending on the institutional environment, there are two distinct patterns of controlling relationships, where trust and power are inter-related in significantly different ways. First, both mechanisms are generated at the inter-personal level and either trust or power dominates the relationship. Second, power occurs at the level of the institutional framework of relationships and then feeds into the process of developing trust between organizations. In this paper, specific forms of trust and power are identified and the structures of the socio-economic environment is viewed as playing a crucial role in shaping the quality of trans-organizational relations. The theoretical argument mainly draws on conceptual ideas of Systems Theory, Structuration Theory and New Institutionalism. Empirical illustrations are taken from a comparative study in the quality of German and British supplier relationships in the mining machinery and kitchen furniture industries. 

 

Implementing Trust-Based Decision Support Systems
Christian Jensen, Technical University of Denmark 
Trust may be used to support decision making in different ways; either to choose between multiple potential partners for a given transaction or to whether a potential partner is sufficiently trustworthy to allow the interaction to proceed. In either case, a number challenging issues have to be addressed by a trust-based decision support system definition of an appropriate trust model, including policies for trust formation, trust evolution and trust exploitation, definition of a computational representation of trust, the relationship between trust and risk, and the impact of context of the proposed interactio on the trusting relationshp between the different parties.

Based on our experiences from the SECURE project, which is developing a trust-based security framework, we examine these issues, identify different approaches to their implementation and present some of the solutions proposed in the SECURE project.

 


Risk Management
Berc Rustem, Imperial College London 

We discuss two complementary approaches to risk management. Both are based on the characterisation of the future in terms of scenarios. One is probabilistic, the other is a worst-case approach. The underlying concept in both is optimal decision making. We discuss the trade-off between the conflicting objectives of performance and the associated risk.

Trust and Roles
Andrew Jones, Kings College London, UK 
One of the most common uses of the term 'trust' occurs in connection with agents who occupy particular roles. For instance, when one speaks of trusting one's doctor, or not trusting the Prime Minister. This presentation will outline a characterisation of the concept of role, together with an analysis of what it means to trust a role-occupant.



Ethics and Trust Intelligent Trust, Intelligent Accountability
Onora O'Neill, University of Cambridge, UK
Current public debates often contrast or even oppose trust and accountability. The contrast is typically made plausible by concentrating on particular conceptions of each. The favoured conceptions of trust see it as a matter of attitude and affect, and emphasise how risky trust can be even within, let alone beyond, personal relations they leave little room for a conception of reasonable trust. The favoured conceptions of accountability are managerial. They identify accountability with procedures by which institutions and managers set targets, check performance against those targets, and sanction poor performance, and are closely linked to (aspirations to) managerial control.

Few reasons are given for favouring these conceptions of trust and accountability, or for rejecting or neglecting others. Consideration of a range of conceptions both of trust and of accountability suggests that broader and more intelligent conceptions of each are available. These conceptions are preferable because they recognise the propositional character of trusting and of rendering an account; because they allow for the role of testimony and judgement in placing trust and rendering an account to others; and because they show how intelligent forms of accountability support rather than supersede intelligent trust.