# Expression of Interest

## Name of Project Lead and Co-Leads

#### Alan Turing Institute
- Christopher Burr (Project Lead)
- Jennifer Ding (Project Co-Lead)
- Research Software Engineer (DI)
- Sophie Arana (DA)
- Research Community Manager (DI)
- Research Associate (DI)
- Research Project Manager (DI)

#### University of York
- Ibrahim Habli (Project Co-Lead)
- Phillip Morgan (Project Co-Lead)
- Nathan Hughes (Research Associate)
- Stefano Faraoni (Research Associate)

## Host Research Organisation (RO) and other Research Organisations involved in the project

- Alan Turing Institute (Host)
- University of York

## Name of non-academic partners

- British Antarctic Survey
- Met Office
- Heartflow
- NHS England
- Digital Twin Hub (Innovate UK)
- Pinsent Masons
- Microsoft
- Google DeepMind

## Approximate Full Economic Cost (fEC) of the project

- £1.1 million

## Approximate Combined Cost of Co-Investigators not based at eligible research institutions

- Detail 1

## Which research discipline within UKRI remits will your project cover

- Philosophy & Ethics
- Law
- Human-Computer Interaction
- Software Design

## Total Number of Researchers from an AHRC Discipline

5

## Summary

Current policy has emphasised the strategic importance of building resilient and human-centred cyber-physical infrastructure as a vital means for securing a sustainable and healthy society (e.g. mitigating adverse effects of climate hazards on public health and critical infrastructure). However, there is significant conceptual, scientific, and technological uncertainty that needs to be addressed to ensure that any proposed solutions are fair and aligned with the public good.

Building on our scoping research, funded through the BRAID programme, we will work with existing and new project partners to demonstrate how open tools and practices for assuring cyber-physical infrastructure (e.g. connected digital twins and other AI-enabled systems) can help reduce the associated uncertainty inherent in building resilient and human-centred infrastructure. To do this, we will draw from and subsequently develop existing research from across the Arts and Humanities (e.g. philosophy, law, human-computer interaction). For example, we will show how philosophical approaches to argumentation (e.g. informal logic and argumentation theory) can underpin novel approaches to assurance and make them more human-centred (i.e. explainable), and how legal frameworks can be made more resilient by carrying out wide-ranging "penetration testing" in sandbox environments but with real-world examples.

With this additional funding, we will secure the impact and sustainability of our initial scoping research and enhance the usability of our open-source and community-centred infrastructure, known as the Trustworthy and Ethical Assurance platform. For instance, we will expand our current repository of assurance case studies, being co-developed with project partners, to include several more extensive investigations of assurance mechanisms across complex project lifecycles and data-driven pipelines. This will widen our structured knowledge base, allowing for new members of the assurance ecosystem to benefit from and contribute to our open and co-produced materials.

In line with the requirements of this call, we will focus on two of the BRAID pillars, *resilient* and *human-centred* innovation, within the domain of *cyber-physical infrastructure* (e.g. explainable and human-centred interfaces for digital twins as decision support systems). More specifically, we continue to focus on multi-disciplinary use cases for digital twins and AI systems across *health*, *natural environment*, and *physical infrastructure*, by bringing together projects from our own research institutions. For this call, our use cases will include:

1. A project using patient and population data, informed by physiology and physics-based models and AI tools, to create a digital twin healthcare pathway.
2. A project developing a synthetic data pipeline to allow policy-makers and decision-makers to determine the effects of climate hazards on vulnerable populations.
3. A project developing a digital twin of a building to optimise efficient energy use.

These projects will also be enhanced through strategic knowledge sharing activities with our project partners and wider community, who will widen our stakeholder and user engagement activities and also provide non-academic examples to improve our evaluation and validation (e.g. DT Hub, Met Office, NHS England, Pinsent Masons, Google DeepMind).

## Objectives

1. Co-create open, community-centred tools, practices and systems for assuring cyber-physical infrastructure (e.g. connected digital twins).
2. Carry out inclusive and diver multi-stakeholder engagement to build capabilities within the UK's emerging AI assurance ecosystem.
3. Demonstrate impact and value of responsible research and innovation practices, facilitated through the TEA platform, and use gathered empirical evidence to co-produce additive best practices and standards.
4. Produce high-quality open and reproducible outputs and deliverables that demonstrate the value of open science and software practices and models, and further enable sustainable communities of practice for assuring cyber-physical infrastructure.