AI in healthcare might spark visions of robotic doctors and futuristic diagnostics. But Robbie Jerrom, senior principal technologist for AI at Red Hat, insists the real gains lie in the mundane systems that make the NHS run more efficiently.
“Prioritising summarising email requests or patient call logs… they’re not very exciting, but they consume a lot of time for staff”, he told City AM. “Small productivity gains really do add up”.
Jerrom argued that pressures on frontline staff, particularly GPs, make the potential of AI transformative.
“With generative AI systems, they have a companion to help keep them updated”, he said.
“Ultimate responsibility is still down to the GP or medical professional, but it gives them access to a lot more knowledge… and fast time-frames matter. They save lives”.
Real world examples
AI’s benefits are already being tested overseas. Jerrom highlighted a US Veteran Affairs programme where AI tools flagged early signs of suicide risk from social media and correspondence data.
“We were able to look at social media feeds… and identify patients that were at risk, and then be proactive, get out and get them helping”.
“And that’s gone from a proof of concept hackathon to moving into a full pilot so they can actually catch these things before we read about it in the news, or before someone has unfortunate and takes their own life”.
He suggested similar approaches could detect early-onset dementia or mental health decline in the UK, “it’s really about finding those patterns”.
Even more routine tools can free up staff time. Voice agents, call transcription, and AI-driven triage could reduce repetitive work and route patients to the right help faster.
“AI has the capability to be a voice agent that you can call and you can ask for help”, Jerrom explained.
Data, privacy and sovereign AI
The NHS holds one of the largest patient datasets in the world, a potential goldmine for AI, but one that demands careful handling.
“Pulling that information together and making sure that is secure, safe and you aren’t sharing confidential information across systems is always a challenge”,Jerrom said. “Governance of confidentiality is as critical”.
He also emphasised the importance of keeping AI systems transparent and homegrown where sensitive data is involved:
“With medical data, with actual UK citizens’ medical information, keeping that safe and secure and having that information processed on an AI system that is sovereign is essential. We can’t have a black box we just can’t look into”.
Open-source AI allows visibility into how models are trained, ensuring systems aren’t relying on opaque or unreliable data.
The bigger picture
Jerrom frames AI not as a silver bullet, but as a practical tool that complements human judgement, manages NHS workload, and ensures patients get timely care.
Whether through routine task automation, data-driven insights, or predictive analytics, he argues, it is the background, “boring” applications that will save lives.
“Small gains add up… tasks like summarising call logs, or using voice assistants for basic interactions consume huge amounts of staff time”.
“When AI takes on those jobs, you suddenly free up professionals to spend more time with patients.”