ECSCW 2026 Workshop

Workshop Schedule   |   Workshop Contributions

Towards Trust Architectures and Assurance of AI-based Systems

This workshop sets out to explore our understanding of the mechanisms underpinning interpersonal trust, institutional trust and trust in technology. In particular, it seeks to explore how these different kinds of trust complement and interrelate with one another to form a ‘trust architecture’ that supports effective and dependable use of technologies and tools. With the rapid push to adopt AI-based systems within organisations in both the private and public sectors, there is an urgent need to determine what actions – both social and technical – should be taken to establish and adapt this trust architecture. This is central to developing trust in AI-based systems and assurance and governance processes that can then sustain that trust. We invite position papers, empirical studies, requirements investigations, and potential solutions relating to these concerns.

Background

In recent years, the literature regarding trust in digital systems has begun to engage with “the social mechanisms which generate trust” (Luhmann, 1990: 95). There is little agreement, however, on what trustworthiness as applied to technology and technology use actually involves. The importance of understanding trust relationships – interpersonal trust, institutional trust and trust in technology – is becoming increasingly important as the pace of AI-based systems adoption accelerates and they are being used in roles that are sometimes radically different from those occupied by previous generations of digital systems. Chander et al. (2025) outline a number of requirements for trust in AI-based systems: transparency, including traceability, explainability and communication; technical robustness and safety, including resilience and security; and accountability, including auditability. Kaur et al. (2022) examined various approaches for addressing trust in AI-based systems, including fairness, explainability, accountability and privacy. Alzubaidi et al. (2023) proposed accuracy, reproducibility, robustness, human agency and oversight. Various other studies, surveys and frameworks have come to similar conclusions and recommendations (Berman et al., 2024; Bagheri & Dirksen, 2024; Ferdaus et al., 2026). Of particular relevance for this workshop are perspectives on trust and trustworthiness that differentiate between ‘trust in technology’ (Lee & See, 2004), interpersonal trust and ‘institutional trust’ – the organisational work settings within which technology gets used (Geffen et al., 2006). Trust in technology is then a socio-material relation between people, their work practices and tools that is both ‘constitutive’ of, and resultant from forms of interaction and organisational work (Watson, 2009).

Our own work reveals that trust in technology is not a binary relation but has contours that mark out regions of higher or lower trust (Procter et al., 2023; Procter & Rouncefield, 2025). If AI-based systems are to be embedded successfully within organisational decision-making routines then it is essential to understand: how trust is calibrated, interrogated and its gradients navigated as a “mundane feature of everyday work” (Clarke et al., 2007: 4) and ‘ordinary action’ (Voss et al., 2006); the roles of interpersonal trust, institutional trust and trust in technology as constitutive elements of a ‘trust architecture’; and how organisations should act if they are to adapt these trust architectures, assurance and governance processes to establish and then sustain trust in AI-based systems.

Workshop Themes

Our aim for this workshop is to gain a better understanding of how trust architectures in different work settings sustain trust and the challenges of adapting them as AI-based systems become progressively embedded in organisational work practices and workflows. AI assurance and governance are now recognised as major problems, especially for high-reliability organisations involved in safety-critical application domains, such as healthcare (DSIT, 2024; NIST, 2023) but are also registering as a more general issue as businesses seek to exploit AI to improve their productivity (Niederhoffer et al., 2025). Our workshop will explore how organisational trust architectures, assurance and governance procedures and processes can be adapted so that the trustworthiness of AI-based systems can be monitored and audited, and what kinds of actions might then be taken when problems are detected.

AI-based tools and systems are increasingly being adopted across a broad range of both the private and public sectors. However, little is currently understood about the challenges associated with ensuring that AI-based tools can be employed productively – and in some application domains, safely – and how to address them. What is becoming clear, however, is that one important factor is the kinds of trust relationships that can be established between users and AI-based systems, and what kinds of procedures and processes for assurance and governance are required if that trust is to be sustained. This workshop sets out to explore our understanding of the mechanisms underpinning interpersonal trust, institutional trust and trust in technology, how these are constitutive of a ‘trust architecture’, and what steps – both social and technical – may be taken if organisations are to adapt their existing trust architectures, and assurance and governance procedures and processes to establish and sustain trust in AI-based systems. 

Call for Participation

We invite a range of contributions along the following lines:

Position papers: What is currently known about the challenges and solutions to trust in AI-based systems as their use proliferates within a wide and diverse range of applications and decision-making practices? What is AI assurance and what practical forms might this take within organisations that are or aim to embed AI-based systems within their workflows?
Empirical studies: What lessons are there from studies of trust relationships with digital systems that would help inform AI-based systems? What kinds of processes do high reliability organisations rely on to assure robustness and quality? How do teamwork and collaboration contribute to assurance processes that sustain trust in technologies?
Requirements: What kinds of actionable insights for establishing and sustaining trust in AI-based systems can we draw from such studies?
Potential solutions: How can the challenges of trust, assurance and governance of AI-based systems be addressed, for example, through the use of techniques to enhance AI-based system transparency? How should organisational assurance procedures and processes be adapted for the roles that AI-based systems are or will be expected to occupy within existing or new workflows?

Submissions should be between 5-8 pages in length + references using the recent EUSSET Conference Template. Submissions should be sent to ai-workshop-ecscw26@wineme.wiwi.uni-siegen.de.

Authors of accepted papers will be invited to submit longer versions to be included in a special issue of a journal. The workshop organisers aim to have an update on this at the conference.

Workshop Activities and Goals

Workshop activities will include:

Morning 09.30-12.30

  1. Opening keynote (20 minutes).
  2. Presentations (3) of accepted papers (20 minutes per paper, 5 minutes discussion).
  3. Coffee break.
  4. Presentations (3) of accepted papers (20 minutes per paper, 5 minutes discussion).

 Afternoon 14.00-16.30

  1. Presentations (3) of accepted papers (20 minutes per paper, 5 minutes discussion).
  2. Coffee break.
  3. Panel discussion.
  4. Concluding remarks.

The goal will be to explore through position papers and empirical studies the challenges of devising solutions to the challenges of embedding AI-based systems safely and effectively in a range of organisational settings and application domains.

Important Dates

  • Position Paper Submissions to our Workshop: 30 May 2026 (AoE)
  • Workshop Day: Tue, 30 June 2026

Workshop Organizers

Rob Procter is a Professor in the Department of Computer Science at the University of Warwick and a fellow of the Alan Turing Institute for Data Science and AI. His principal research interest is social informatics, understanding how individual, organisational and social factors shape processes of appropriation (design, development and adoption) of digital innovations. He has conducted research in a wide range of organisational settings, including healthcare, financial services, industry and the public sector. Procter specialises in the use of ethnographic methods to understand the implications of digital innovations for work practices (e.g., decision-making, collaboration, coordination) and how organisations respond to the challenges of adoption. Much of his work is action research-oriented, using insights from Science and Technology Studies and Social Shaping of Technologies to understand and facilitate innovation processes. His current research includes studies of trust and assurance of AI-based systems in air traffic control, clinical practice and law.

Mark Rouncefield is a Senior Professor at the University of Siegen, Germany and formerly a Reader in Social Informatics in the School of Computing and Communications, Lancaster University. His research interests are in Computer Supported Cooperative Work and Computer Human Interaction and involve the study of various aspects of the empirical study of work, organisation, human factors and interactive computer systems design, working across traditional disciplinary boundaries to address challenging socio-technical problems. He is particularly associated with the development of ethnography as a method for informing design and evaluation and his empirical studies of work and technology have contributed to important debates concerning the relationship between social and technical aspects of digital systems design and use.

Peter Tolmie is a Principal Research Scientist in the Information Systems and New Media group at the University of Siegen in Germany. He has conducted a wide range of ethnographic studies across numerous domains, including small businesses, home environments, artistic experiences, musical performance and production, secretarial work, museums and galleries, the TV and film industries, bid management, healthcare, and journalism. He has been published widely in both journals and conferences in the domains of Computer-Supported Cooperative Work and Human-Computer Interaction and is the author and editor of a number of books relating to ethnomethodological studies of various kinds.
 
Dr. Claudia Müller is a full professor at the University of Siegen, where she holds the Chair of Information Systems, specifically focusing on IT for the Ageing Society. Her work is highly interdisciplinary, blending computer science with gerontology and social sciences. Her research focuses on Socio-Informatics, Digitalisation, and Participatory Design to support older adults in maintaining social inclusion and health. She is a partner in the Praxslabs initiative, using practice-based methodologies to ensure technology meets the real-world needs of an ageing population. She served as the deputy chairwoman for the German Federal Government’s Eighth Age Report (Achter Altersbericht). She is a speaker for the Gerontological Research Network (GeNeSi) and a member of the “Age & Technology” working group within the German Society of Gerontology and Geriatrics.