In the latest of a string of reports issued over the last 12 months by bodies in and around UK Government on Artificial Intelligence (AI), the Parliamentary Office of Science and Technology (POST) has just published a research briefing on AI and healthcare.
POST is a non-partisan research body, so the report has no immediate impact on current or planned legislation, but it does provide some possible signposts for AI suppliers on what a post-Brexit AI regulatory regime will look to address.
Key takeaways include:
- an acknowledgement that AI has the potential to revolutionise the way in which healthcare is delivered, its cost, and consistency in patient diagnosis and outcomes;
- recognition that the scope and volume of datasets underpinning the AI will need to be vast, an issue that is likely to challenge both AI developers and public trust in the way in which their personal data is used;
- the particular risk in a healthcare context of AI systems producing ‘algorithmic bias’, resulting in recommendations that discriminate against certain demographics; and
- questions of how liability would be fairly allocated where AI or AI-based decisions cause patient harm.
In many ways the issues raised don't really differ from those generally associated with AI. The possible difference here is that, by their nature, healthcare settings provide a context for stress-testing the ethical and technical limits involved in the rollout of AI in 'real-world' scenarios.
Bodies like NHS-X and its NHS AI Lab are pushing ahead on the application of digital technologies, including AI, in healthcare but the question now is whether a coherent legislative regime will emerge to support this.
Following conclusion of the Brexit withdrawal terms in December, the Government arguably has greater room for manoeuvre on future regulation - and has already opted out of implementing the Medical Devices Regulation which will come into effect in EU countries in May. As with many other policy areas, however, that new-found freedom will only sharpen the focus on what the Government intends to do to fill the gap.
AI has the potential to improve health outcomes and offer cost savings through reducing the time spent by staff on routine work