Clinical AI Governance
“Our Clinical Artificial Intelligence Governance is formed of seven pillars that merge statutory, regulatory and professional best-practice to ensure our AI helps make it easier for our users to take care of themselves and their loved ones.”
Dr Keith Grimes, Clinical AI Director
At Babylon we believe everyone, everywhere, has the right to a long, healthy and happy life. With medical professionals in short supply we are using artificial intelligence (AI) to amplify the impact of their work.
We use a robust system of governance to transparently measure and improve the AI and ensure it is used as appropriately as possible. We also adhere to the United Kingdom’s code of conduct for artificial intelligence (AI) systems used by the NHS.
Our approach to AI Governance follows the NHS guidance of:
Quality Audit
We have a cyclical process of improving the quality of our AI. We review the performance of our AI, compare it against standards and refine it as a result. These standards may be internally set, based on historical performance, developed through expert consultation and adaptation of existing clinical standards, or come from the growing range of national and international standards of best practice.
Effectiveness
We apply knowledge derived from our Scientific Research team, external research, clinical experience and user studies, to keep improving outcomes for users whilst optimising processes. We monitor practice, inform our teams and implement changes.
Education and Training
We educate and continually train not only the staff who develop and test the AI, but also the AI itself. This helps ensure the highest quality of clinical data and knowledge is used in the development of all our AI, and that every effort is made to ensure that conscious and unconscious bias is minimised.
Research and Development
Our Scientific Research team publishes peer-reviewed research and works with researchers from multiple universities to ensure best research practice. We are working with the World Health Organisation and others to establish a standardised benchmark for symptom checkers.
Risk Management
Our robust methodology assesses and manages risk to users, staff and the healthcare system, identifying and responding to incidents, complying with safety standards, and operating to the highest ethical standards. This includes management of any internal or external reports received. In assessing and managing risk, we comply with the NHS’ DCB0129 and DCB0160 Clinical Risk Management Standards in the UK, adhere to the National Institute for Health and Care Excellence (NICE) Guidance on the ‘Evidence Standards Framework for Digital Health Technologies’, utilise the NHS Clinical Risk Matrix throughout our testing and assurance processes, and globally apply EN ISO 14971:2012 risk management to our medical devices.
Information Management
We comply with applicable data protection laws (including the GDPR) and follow best practice in the ethical, secure, and legal use of patient data in the delivery of healthcare, whilst ensuring that the full value of this important resource is realised for the greatest benefit of the patients.
Openness
We engage with the global community of clinicians, our users and the wider public and welcome both feedback and compliments! We use extensive user research to guide changes, our users can feedback instantly using a real-time star rating system and we have multiple easy ways of getting in contact. All our staff can raise any concerns and we work closely with trusted bodies across the world including: World Health Organisation, NHSX, NHS Digital and leading academics in the US, UK and China.