AI technologies have the potential to unlock significant growth for the UK financial services sector through novel personalised products and services, improved cost-efficiency, increased consumer confidence, and more effective management of financial, systemic, and security risks. However, there are currently significant barriers to adoption of these technologies, which stem from a capability deficit in translating high-level principles (of which there is an abundance) concerning trustworthy design, development and deployment of AI technologies ("trustworthy AI"), including safety, fairness, privacy-awareness, security, transparency, accountability, robustness and resilience, to concrete engineering, governance, and commercial practice. In developing an actionable framework for trustworthy AI, the major research challenge that needs to be overcome lies in resolving the tensions and tradeoffs which inevitably arise between all these aspects when considering specific application settings.For example, reducing systemic risk may require data sharing that creates security risks; testing algorithms for fairness may require gathering more sensitive personal data; increasing the accuracy of predictive models may pose threats to fair treatment of customers; improved transparency may open systems up to being "gamed" by adversarial actors, creating vulnerabilities to system-wide risks. This comes with a business challenge to match. Financial service providers that are adopting AI approaches will experience a profound transformation in key areas of business as customer engagement, risk, decisioning, compliance and other functions transition to largely data-driven and algorithmically mediated processes that involve less and less human oversight. Yet, adapting current innovation, governance, partnership and stakeholder relation management practice in response to these changes can only be successfully achieved once assurances can be confidently given regarding the trustworthiness of target AI applications. Our research hypothesis is based on recognising the close interplay between these research and business challenges: Notions of trustworthiness in AI can only be operationalised sufficiently to provide necessary assurances in a concrete business setting that generates specific requirements to drive fundamental research into practical solutions, with solutions which balance all of these potentially conflicting requirements simultaneously. Recognising the importance of close industry-academia collaboration to enable responsible innovation in this area, the partnership will embark on a systematic programme of industrially-driven interdisciplinary research, building on the strength of the existing Turing-HSBC partnership. It will achieve a step change in terms of the ability of financial service providers to enable trustworthy data-driven decision making while enhancing their resilience, accountability and operational robustness using AI by improving our understanding of sequential data-driven decision making, privacy- and security- enhancing technologies, methods to balance ethical, commercial, and regulatory requirements, the connection between micro- and macro-level risk, validation and certification methods for AI models, and synthetic data generation. To help drive innovation across the industry in a safe way which will help establish the appropriate regulatory and governance framework, and a common "sandbox" environment to enable experimentation with emerging solutions and to test their viability in a real-world business context. This will also provide the cornerstone for impact anticipation and continual stakeholder engagement in the spirit of responsible research and innovation.

This proposal focuses on the electricity network of 2050. In the move to a decarbonised energy network the heat and transport sectors will be fully integrated into the electricity system. Therefore, the grand challenge in energy networks is to deliver the fundamental changes in the electrical power system that will support this transition, without being constrained by the current infrastructure, operational rules, market structure, regulations, and design guidelines. The drivers that will shape the 2050 electricity network 2050 are numerous: increasing energy prices; increased variability in the availability of generation; reduced system inertia; increased utilisation due to growth of loads such as electric vehicles and heat pumps; electric vehicles as randomly roving loads and energy storage; increased levels of distributed generation; more diverse range of energy sources contributing to electricity generation; and increased customer participation. These changes mean that the energy networks of the future will be far more difficult to manage and design than those of today, for technical, social and commercial reasons. In order to cater for this complexity, future energy networks must be organised to provide increased flexibility and controllability through the provision of appropriate real time decision-making techniques. These techniques must coordinate the simultaneous operation of a large number of diverse components and functions, including storage devices, demand side actions, network topology, data management, electricity markets, electric vehicle charging regimes, dynamic ratings systems, distributed generation, network power flow management, fault level management, supply restoration and fuel choice. Additionally, future flexible grids will present many more options for energy trading philosophies and investment decisions. The risks and implications associated with these decisions and the real-time control of the networks will be harder to identify and quantify due to the increased uncertainty and complexity.We propose the design of an autonomic power system for 2050 as the grand challenge to be investigated. This draws upon the computer science community's vision of autonomic computing and extends it into the electricity network. The concept is based on biological autonomic systems that set high-level goals but delegate the decision making on how to achieve them to the lower level intelligence. No centralised control is evident, and behaviour often emerges from low-level interactions. This allows highly complex systems to achieve real-time and just-in-time optimisation of operations. We believe that this approach will be required to manage the complex trans-national power system of 2050 with many millions of active devices. The autonomic power system will be self-configuring, self-healing, self-optimising and self-protecting. This proposal is not focused on the application of established autonomic computing techniques to power systems (as they don't exist) but the design of an autonomic power system, which relies on distributed intelligence and localised goal setting. This is a significant step forward from the current Smart Grid vision and roadmaps. The autonomic power system is a completely integrated and distributed control system which self-manages and optimises all network operational decisions in real time. To deliver this, fundamental research is required to determine the level of distributed control achievable (or the balance between distributed, centralised, and hierarchical controls) and its impact on investment decisions, resilience, risk and control of a transnational interconnected electricity network. The research within the programme is ambitious and challenges many current philosophies and design approaches. It is also multi-disciplinary, and will foster cross-fertilisation between power systems, complexity science, computer science, mathematics, economics and social sciences.