Lords Committee: Artificial Intelligence: Is it good for our health?

The Committee will hear evidence on how AI is already used in healthcare, its potential future uses, whether the NHS has the capacity to take advantages of the opportunities represented by AI technology, and what ethical standards may be potentially required in the development and deployment of AI in the UK.

These are among the issues the House of Lords Select Committee on Artificial Intelligence will be raising with two panels of witnesses on Tuesday 21 November 2017 from 3.30pm.

The first panel will be at 3.30pm and the Committee will hear from:

• Dr Julian Huppert, Chair, Independent Review Panel for DeepMind Health
  The review panel scrutinises the work of DeepMind Health, a branch of DeepMind, the world leader in artificial intelligence research. Dr Huppert is a former MP and his research focuses on science and technology policy.
• Nicola Perrin, Head, Understanding Patient Data
  Understanding Patient Data is an initiative from the Wellcome Trust that is designed to support better conversations about the use of data in healthcare. Its website aims to provide objective information about the way data is used, with the explanations given in such a way that everyone can understand, regardless of their medical or data science knowledge.
• Dr Sobia Raza, Head of Science, PHG Foundation
  The PHG Foundation is a health policy think tank committed to helping decision makers deliver the benefits of biomedical innovations to all. Dr Raza is responsible for the delivery of the scientific expertise and analysis that underpins its work.

Questions the Committee is likely to ask include:

• How does the public view the use of AI in healthcare?
• How should the liability and compensation for any errors made by AI applications be handled?
• Would making all publicly-generated health data publicly available, subject to anonymisation, encourage progress in AI research and innovation?
• Who should own algorithms developed using NHS data?
• Does the NHS have the capacity to maximise the opportunities and minimise the risks represented by AI technology?
• What should the NHS look for when buying AI applications that have been developed in specific contexts using unrepresentative datasets?

The second panel will be at 4.30pm and the Committee will hear from:

• Dame Fiona Caldicott, National Data Guardian for Health and Care, Office of the National Data Guardian
  The National Data Guardian helps to ensure that personal confidential data is held and used to support better outcomes from health and care services, at the same time providing confidence that there are thorough safeguards in place to protect personal confidential data. Dame Fiona was appointed to this role in November 2014. 
• Dr Hugh Harvey, Clinical Artificial Intelligence Researcher and Consultant Radiologist, Guy’s and St Thomas’ NHS Foundation Trust
  Last year Dr Harvey spent a year at Babylon Health, a digital health and artificial intelligence start-up, working in clinical artificial intelligence and predictive analytics. Dr Harvey was the lead for regulatory affairs in clinical artificial intelligence, overseeing the regulatory affairs regarding stand-alone software as a medical device. 
• Professor Martin Severs, Medical Director, NHS Digital
  NHS Digital is an executive non-departmental public body which uses information and technology to improve health and care. As a Caldicott Guardian, Professor Severs is central to ensuring that NHS Digital’s information governance structures and processes are of best practice and transparent.

Questions the Committee is likely to ask include:

• Can data ever be truly anonymised?
• Could the creation of data trusts or a data stewardship body effectively broker safe access and ethical use of data?
• In the deal between DeepMind and the Royal Free NHS trust, how was the value of 1.6 million patient records determined? Has the NHS retained any rights over the applications, such as the Streams app, that have been developed as a result?
• What level of transparency should be insisted on when using AI in a clinical context?
• How effective is the use of chatbots and virtual medical assistants to triage patients in healthcare?

These evidence sessions will take place at 3.30pm on Tuesday 21 November 2017 in Committee Room 4A of the House of Lords.