AI is not a “cure-all solution” for the NHS, expert warns

Artificial intelligence (AI) can be “transformative” for healthcare, but still faces “lingering doubts”, InnoScot Health has said.

NHS Scotland's partner has said there is “lot of hard work to be done" for NHS workers to trust the technology is safe, ethical, and does not threaten jobs.

Executive chair of InnoScot Health Graham Watson, has urged for more real-world evidence to prove “beyond doubt” the long-term potential of the cutting-edge technology in health and care systems.

Watson said: “Arguably, healthcare with its rich and growing wellspring of data drawn from the likes of clinical studies, imaging, and wearables is the perfect use case for AI when compared to other sectors. The potential is vast, but there remain risk factors.

“The excitement around it must also be tempered in our current challenging climate of renewal and transformation where human expertise is still paramount and to which any technology must remain only supplementary.”

Watson's comments come after a survey showed NHS staff are divided on the benefits AI could bring to their work.

Graham Watson | InnoScot Health

Gathering more than 1,200 responses, a survey by the Health Foundation found a similar number of workers said their jobs would be mostly threatened or improved by AI, with midwives, nurses and staff in administrative and clerical roles less likely to welcome it compared to doctors and other health professionals.

Meanwhile, more than two-thirds of those surveyed said AI would make them feel “more distant” from patients.

However, a majority – 57 per cent – said they look forward to using AI as part of their jobs, making it “clear” AI has also gained “significant support”, Watson said.

He continued: “AI may well become a large piece in tomorrow’s healthcare innovation puzzle, but it should not be considered a cure-all solution, particularly so when health boards differ in the infrastructure available to host its capabilities and, by extension, their approach to both integrating it and successfully delivering bespoke training for staff.

“If or when AI technologies gain workforce, patient, and regulatory approval, adoption could furthermore be slow, meaning that expectations must be managed in the short-term.

“Neither should anyone sideline concerns of NHS staff and patients around AI’s ethics, accuracy of decision making, or care becoming less human-led.”

He continued: “For decision makers considering these survey results, it appears that the successful introduction of AI into healthcare will require human presence to remain to the fore, and fundamentally its enabling technology must deliver proven results in order to command both patient and staff confidence.”