Trustworthy digital infrastructure for identity systems – Impact projects

Introduction

The Turing is commissioning a series of discrete, 6-12 month impact projects that complement or enhance the depth of the core research effort. A first series of seven impact projects from five leading UK universities explore varied themes and field insights, from attitudes toward privacy and institutional trust to advancements in the analysis of stochastic processes, challenges faced by home and migrant workers, and threat modelling for comparative risk in self-sovereign ID systems.

A second call for contributions will be extended in the second half of 2021.

University of Warwick

Speak for Yourself!

Attitudes to contact tracing applications in the context of COVID-19: results from a nationally representative survey of the UK population to informing governments of appropriate design choices for adequate uptake and participation.

Click to read more

Title: Speak For Yourself!

Team: Professor Carsten Maple (PI), Dr Rebecca McDonald

Description: To undertake a survey on the design of contact tracing apps. This proposal involves undertaking a nationally representative survey to elicit views of trust in an NHS contact tracing App.

Explaining the science: There is an ongoing debate in the public arena about the use of app-based contact tracing to help manage the COVID-19 pandemic. A number of countries have deployed contact tracing techniques to address the spread of the disease. A trial of a centralised UK app is ongoing on the Isle of Wight.

Despite controversy around what approach is in the public’s best interest, as yet, the opinions of the public have not been gathered, analysed or considered at a representative scale. We have undertaken a nationally-representative survey of the UK public. We utilised a specific method, a Discrete Choice Experiment (DCE), to help understand public opinion on aspects of contact tracing apps.

Real world application: The purpose of this work is to help inform those intending to design and deploy contact tracing apps in time of pandemic, allowing governments to make appropriate design choices to ensure adequate uptake and participation. In cases where concern exists, but a government has an overriding requirement, the insights can inform awareness and informational campaigns, to increase understanding of the design choice.

Read full report

Mechanistic marked spatio-temporal point processes

Exploring event- sequence data, one of the most abundant data structures in both natural and artificial ecosystems, resulting from phenomena as diverse as earthquakes, disease outbreaks, economic cycles, and social processes within digital social networks for a novel mechanistic modelling framework.

Click to read more

Title: Mechanistic marked spatio-temporal point processes for large-scale data-analytic applications in identity systems and cyber-security.

Team: Dr Ioannis Kosmidis (PI), Professor Petros Dellaportas (Co-I), Dr Aristeidis Panos (PDRA) 

Description: Event-sequence data is one of the most abundant data structures in both natural and artificial ecosystems, resulting from phenomena as diverse as earthquakes, disease outbreaks, economic cycles, and social processes within digital social networks. The goal of this research project is to revolutionise the analysis of event-sequence data through a novel mechanistic modelling framework for marked spatio-temporal point processes.

Explaining the science: Digital identity services and systems give rise to multiple, multilevel event sequences of unprecedented frequency and detail, both at the user level (e.g sharing of personal information and interaction with social media platforms, use of biometric devices for purchases, authentication, identity verification, and so on), and at the service/system level (such as connection requests between nodes, authentications, ad-hoc communication between linked services, and so on). Such event sequences also come with a wealth of accompanying event- or process-specific information (e.g. user/system/service characteristics, text, location, images, videos, and so on) which is now routinely recorded.

While it is widely recognised that the statistical handling of such data can lead to valuable insights with significant impact on the operation and development of identity systems and to seminal contributions in dynamic cyber-security, current statistical modelling frameworks are challenged by the data's variety, volume, velocity and heterogeneity. The goals of the proposed research program are to:

revolutionise the analysis of event-sequence data through a novel mechanistic modelling framework for marked spatio-temporal point processes (MSTPPs);
deliver algorithms of realistic complexity for online and large scale statistical learning of interpretable characteristics of MSTPP, along with inferential procedures about the local and global effects of accompanying information;
develop open-source software that delivers the developments in G1 and G2 to the data science community;
seek and engage with novel applications, with particular focus on identity systems and cyber-security.

Real world application: Event-sequence data is one of the most abundant data structures in both natural and artificial ecosystems, resulting from phenomena as diverse as earthquakes, disease outbreaks, economic cycles, and social processes within digital social networks. The outcomes of this project addresses these real world challenges.

University College London

UCL logo

Highlighting exclusionary practice

This considers the interface between design and socio-political/economic inequalities to examine the extent to which existing digital and biometric identity systems mitigate barriers to access stemming from class and occupational divides, ethnic and religious marginalisation, gender disparities, generational differences, and citizenship status.

Click to read more

Title: Highlighting exclusionary practices within established and emerging digital and biometric national ID systems in the Global South  Team: Dr Keren Weitzberg (PI)

Description: This project considers the interface between design and socio-political/economic inequalities. It examines the extent to which existing digital and biometric identity systems mitigate barriers to access stemming from class and occupational divides, ethnic and religious marginalisation, gender disparities, generational differences, and citizenship status. Emphasis is placed on the design of digital identity programs and their impact on equity and inclusion.

Explaining the science: This research can offer invaluable insight into the potential risks of exclusion when designing and implementing digital identity systems. It focuses on three cases studies from Brazil, India, and Kenya. Despite their different histories, geographies, and socio-political systems, all three countries share a long legacy of biometric identification dating back to the late nineteenth/early twentieth centuries. In recent years, all three countries have also experimented with and/or implemented centralised, nation-wide biometric registration systems. This report examines why such digital identity systems so often exacerbate problems of exclusion by, as Nikolas Rose suggests, further “policing the obligatory access point” to services. It also poses other questions: to what extent are innovations in ID systems and technical/system design creating novel forms of exclusion or simply mirroring pre-existing inequalities? What recourse do people in different countries and regions have when (not if) exclusion occurs? How might digital identity systems lessen distinctions based on race, ethnicity, gender, or citizenship status? How are biometric systems bringing states, citizens, and residents into new relation with one another? How are they changing the type and nature of the services people can access? What causes people to embrace or reject new biometric identification systems?

Real world application: This research offers insights that can be applied to various settings, from humanitarian/aid to developmental contexts as well as to other national biometric systems.

University of Southampton

University of Southampton logo

Risk models of national identity systems

A knowledge model that represents the forces and factors guiding assessments of trustworthiness from different stakeholder perspectives, and informs the design of signalling mechanisms that communicate trust-relevant information to user communities.

Click to read more

Title: Risk models of national identity systems

Team: Professor Dame Wendy Hall (PI), Professor Michael Boniface (Co-I), Professor Michael Pickering (PDRA), Dr Paul Smart (PDRA)

Description: This project aims to provide a framework that informs the development and evaluation of national identify systems from a trust perspective. The main research output will be a knowledge model that represents the forces and factors guiding assessments of trustworthiness from different stakeholder perspectives. This model will also inform the design of signalling mechanisms that communicate trust-relevant information to user communities.

Explaining the science: The aim of this project is to provide a framework that informs the development and evaluation of national identify systems (NISs) from a trust perspective embedded within a representative human-machine network. Our main research output will be a knowledge model that represents the forces and factors guiding assessments of NIS trustworthiness from the different stakeholder perspectives, who are primary actors within the network. This model will also inform the design of signalling mechanisms that communicate trust- relevant information to user communities as well as provide an initial understanding of end user imperatives associated with explainable AI. Our approach is inherently people-centred, in the sense that trustworthiness will be modelled as the output of a sensemaking process that is performed by people. Nevertheless, our aim is to support the accurate assessment of NIS trustworthiness by representing (and communicating) system-specific technical properties. This will require a consideration of the functional properties of NISs, the broader socio-economic and socio-technical contexts in which such systems are to be deployed, and the disruptive impact of advanced technology implementation within what is currently a trusted environment.

Real world application: Our approach will be validated against two real-world applications: one from within the project, and one external to it. In the first instance, we look at how our definitions of trustworthiness as a construct relate to the project-internal case of piece-meal workers in India to demonstrate the complexities of actor trust-enabling relationships are affected by the introduction of technology. Secondly, and on that basis, we will compare against the UK Government’s contact tracing initiatives. The goal is to provide a framework with which to assess complexities which need to considered when deploying socio-technical solutions into existing networks. As such, the framework echoes similar empirical work within healthcare.

Newcastle University

Verifiable credentials for financial inclusion

Researching the scalability of the identity software for scenarios that include poor connectivity, people movement (travellers, refugees) and large amounts of users, with a focus on human and societal aspects of trust in identity technologies to produce a user-centred design for a scalable system for financial inclusion.

Click to read more

Title: Finclusion: Verifiable Credentials for Financial Inclusion

Team: Professor Aad van Moorsel (PI), Dr Paul Ezhilchelvan (Co-I), Dr Karen Elliott (Co-I), Dr Kovila Coopamootoo (Co-I), Dr Magdalene Ng (PDRA), Dr. Tasos Spiliotopoulos (PDRA), Dr Han Wu (PDRA), Dave Horsfall, Ed Curran

Description: The Finclusion project conducts a stakeholder driven study to determine whether decentralised identities can be effective in improving Financial Inclusion. We implement a scalable software system for financial inclusion using Microsoft's Decentralized Identity and Verifiable Credentials technologies, based on a user-centred design. An important angle in the project's research is to gain understanding of the human and societal aspects of trust in identity technologies for financial inclusion.

Explaining the science: The Finclusion project utilises three main research methods. We use user-oriented design, by talking with stakeholders and take their personal experience and understanding of needs into account. We then design and implement the software system using proven software engineering methods. And finally we do a careful evaluation, both of the technology and the impact on stakeholders and society. The project is highly interdisciplinary, bringing together software engineers, social scientists, psychologists and human-computer interaction experts.

Real world application: The project's real-world application is that of Vulnerability Tokens. Banks and other financial institutions have an obligation to protect vulnerable customers. These vulnerabilities can be far-ranging, from a health-related disability to low language skills and to life events, such as caring responsibilities. There is no satisfactory manner in which these vulnerabilities can be shared with financial institutions. However, if each individual can collect digital tokens testifying to vulnerabilities, then this would provide a means for vulnerability communication and verification. Therefore, decentralised identities and verifiable credentials are an exciting potential solution in this space.

University of Manchester

University of Manchester logo

Resilience in value chain and vulnerability reduction

An in-depth survey of the challenges faced in assigning identities to home and migrant workers with a particular focus in how identity systems can safely support, and empower, extremely vulnerable groups who are incapable of managing their own identities.

Click to read more

Title: Resilience in value chain and worker vulnerability reduction – Trusted digital identity and payments in supply chain.

Team: Professor Ser-Huang Poon (PI), Martin Carpenter (PDRA), Traidcraft Exchange (Subcontractor), Incudeas (Subcontractor)

Description: Informal sector workers are a significant part of manufacturing supply chains yet are subject to various forms of labour exploitation. We are conducting an in-depth survey and stakeholder consultations to contextualise and explore challenges associated with assigning identities to homeworkers, who are typically women and often migrants. In this we are focused on understanding whether and if so, how digital identity systems can support these vulnerable groups and lead to improvement in supply chain transparency and resilience, and ultimately contribute to improvements in working conditions. The resulting findings will inform a high-level design for some proposed technology-assisted solutions intended to achieve these aims.

Explaining the science: Without an identity as a worker, many workers in the informal sector are unrecognised and therefore invisible in the supply chain. They are likely to be subject to very low rates of pay, often below the minimum wage. Lacking an employment contract, they also struggle to access social security and other entitlements under ILO conventions such as freedom of association and collective bargaining. Further there are legal and regulatory risks associated with the use of centralised identity systems for formal sector supply chain actors. Our hypothesis is that these distributed and decentralised patterns of work and manufacture are better served by an identity system that is based on decentralised identifiers and is based on the principles of self-sovereign identity, which mirrors the human patterns of trusted relationships already in operation in the supply chain.

We will use as a case study of the Indian garment manufacturing sector with a particular focus on homeworkers in Kapas Hera in SW Delhi, all of whom are women and migrants. Through Traidcraft India, we will first conduct an in-depth, in-person, survey that illuminates some of the contextual challenges as well as the assignation of identities to workers and supply chain actors. We will also analyse the supply chain mapping flows of product, money and data, and review examples of other digital solutions developed for vulnerable groups subject to similar risks and vulnerabilities. We will then develop potential solutions for a further round of interviews as part of a participatory and inclusive design process. We will particularly consider the needs of users who are unable to manage their own digital identities for example due to low literacy or digital exclusion and factors of access, risk, resilience, and privacy.

The research aims to consult workers from different tiers within the supply chain and garner insights on risks and challenges they face. In doing so, we plan to arrive at feasible solutions and distil major challenges involved in designing such a system. We will also draw out and recommend relevant design principles that any proposed solution for comparable contexts should adhere to.

Real world application: Supporting this work will be a deep focus on homeworkers in the Indian garment sector. The identities of these workers are currently only known to the middlemen who mediate between them and the factories. The precariousness of this position has been forcefully driven home by COVID-19 – their status as workers was unknown leaving industry and the Government less well-equipped to support them at a time of crisis. More generally, an identity method that supported payment would be of great benefit to the homeworkers, and suppliers and brands who are ultimately employing them. Such an ID and payment system must interoperate with existing federated or centralised identity systems.