Health data poverty: what can we do about it?

Dr Xiaoxuan (Xiao, pronounced "Shau") Liu is an Honorary Associate Professor in AI and Digital Health at the University of Birmingham, Principal Scientist and Health Research Lead at Microsoft AI, and a Deputy Editor at NEJM AI.

She was appointed a 125th Anniversary Fellow at the University of Birmingham in August 2024 following her significant contributions to internationally adopted evidence standards and policy for AI and health. Before this, she was an ophthalmology doctor in the NHS and a Health Scientist at Apple.

Dr Liu's work focuses on the responsible innovation of AI health technologies, seeking to ensure they are safe, effective and equitable. Projects she has previously led include:

• Improving the evidence — developing internationally adopted reporting guidelines for AI clinical trials (SPIRIT-AI and CONSORT-AI), and contributing to other AI reporting standards including TRIPOD+AI, STARD-AI and DECIDE-AI; as well as defining standards for NICE (the Digital Health Technologies Evidence Standards Framework), in collaboration with Imperial College London and the Alan Turing Institute.
• Improving safety — developing tools for assessing the safety of AI-enabled medical devices: the Medical Algorithmic Audit and CANAIRI (Collaboration for Translational AI Trials), and working directly with medical device regulators such as the MHRA.
• Improving equity and fairness — tackling algorithmic bias and improving the transparency and diversity of health datasets to mitigate AI-driven health inequalities, through STANDING Together.
• Improving global health — conducting real-world trials of generative AI and large language models to assist community health workers in resource-limited settings, in collaboration with PATH.

Dr Liu works with a range of academic, industry and policy institutions around the world. She supports the UK NIHR Incubator for AI & Digital Healthcare and the NHS AI and Digital Regulatory Services, and serves as a board member for the MHRA's AI Airlock — the regulatory sandbox for AI as a Medical Device.