Automated Ethical AI Assurance Service

Currently there is an explosion of guidance and regulation for the rapidly growing number of companies developing products & services using AI technologies in health. NHS staff and patients are yet to be convinced of the safety of this type of automation, and the rapid pace of change and complexity add to uncertainty. We are going to road-test an approach to ensure AI technologies are designed in an ethical way _before_ any code is written, and their safety and legality reviewed before _and_ during public consumption.

Innovate UK has just announced an online tool intended to provide an easy and faster way to navigate the UK AI ecosystem. The Centre for Data Ethics & Integrity (CDEI) and the Turing Institute are developing principles for trustworthy AI on behalf of the UK Government who will pass them to care quality bodies to enforce. However, digital health innovators are still faced with a confusing choice of frameworks and toolkits and lack practical guidance on how to apply them and understand how they relate to other industry regulations and international standards.

HD-Labs plan to develop a new line of business through rapid, repeated refinement of a prototype toolkit to guide digital health innovators through best practice in ethical AI assurance. This will build on previous research we undertook to help the Organisation for Review of Health & Care Applications (ORCHA) understand how to assess AI technologies in digital health tools that support people to manage their own health. HD-Labs is an existing NHS Supplier with a growing reputation for combining open-source tools to facilitate safe handling and analysis of sensitive data. We have been working with leading academic-industry knowledge transfer initiatives such as the Greater Manchester AI Foundry, NIHR MindTech MedTech Cooperative and the UKRI Trustworthy Autonomous Systems Hub to pioneer adoption of their ethical assurance tools by applying them to our own data orchestration and conversational intelligence products.

This funding will allow us to validate a sustainable business model for ethical assurance of AI technologies that can then be expanded to include those supporting decision-makers in other high-impact contexts. The impact will be to increase trust in and, therefore adoption of, decentralised digital tools to support self-management of long-term conditions. Patients will feel confident and empowered to navigate NHS services and optimise their engagement with clinicians, reducing overall waste.