Speakers: Michelle Williams, University of Edinburgh and Danny Ruta, Guy's Cancer Centre, London
Abstract: Following a series of high profile commercial
and research developments such as alpha go and AlexNet, the 2010s saw a
significant upsurge in attention, interest, and investment in AI technologies.
As part of this much has been written about the potential for AI to
‘revolutionise’ healthcare delivery in diverse domains, from radiography to
mental health. However, the reality of developing systems with concrete and
robust healthcare benefits has proven a significant challenge, as high profile
failures such as IBM’s Watson for Oncology demonstrate. Furthermore, there has
been a significant, and often unappreciated, work to centre the inequitable
impacts and safety risks of AI systems developed when there may have been no
consideration of ethics or societal impact.
To ensure AI systems are
effective, safe, and that substantial steps are taken to reduce and mitigate
for inequitable outcomes, work is now underway to develop methods and
instruments for auditing AI for healthcare. This talk will cover the wider
healthcare contexts and critical issues which establish the need for rigorous
evaluation. The talk will address the role for regulation and standards in
these efforts, with particular attention given to standards. The speakers are
members of the development committee for the UK’s first auditable standard for
Healthcare AI ‘BS 30440 Validation framework for the use of AI in healthcare’.
BS 30440 constitutes an important case study in the opportunities and
challenges presented in developing audit instruments for assessing healthcare
AI.
Bios:
Michelle
Professor Michelle Williams
is a Professor of Cardiovascular Imaging and Radiology Consultant at the
University of Edinburgh and NHS Lothian. She made substantial technical
contributions as a committee member for the BS 30440 development panel. She is
Associate Director of the British Heart Foundation Data Science Centre and the
Imaging theme lead. Her research centers around multi-modality non-invasive
imaging of the heart and blood vessels, including using machine learning and
other advanced analytic techniques. She is president elect of British Society
of Cardiovascular Imaging, member of the executive community of the European
Society of Cardiovascular Radiology, member of the Board of Directors of the
SCCT and chair of the SCCT education committee.
Danny
Danny Ruta is the AI Clinical
Lead at Guy’s Cancer Centre, the cancer centre for Guy's and St
Thomas' NHS Foundation Trust. His present work centres on developing and
applying a validation framework, evidence standards and methods for the clinical
evaluation of AI technology in cancer care. This work is performed
in collaboration with the Clinical Scientific Computing department at
Guy's and St Thomas' and with colleagues at King’s College
London. Danny's draft validation framework formed the foundation
for BS 30440 Validation framework for the use of artificial intelligence
(AI) within healthcare.
Danny has formerly held posts
as Director of Public health for the London Borough of Lewisham and the City of
Newcastle, and as Senior Lecturer in Epidemiology and Health Services Research
at Newcastle University and Dundee University. Danny has developed national
clinical guidelines for the NHS in Scotland, conducted large randomised trials,
and developed Patient Reported Outcome measures recommended by the US FDA, used
with patients across the NHS in England and throughout the world.