AI in Healthcare is increasingly shaping how patients seek help and how clinicians manage demand, as reported by Jamie Smith Webb.
Webb writes that recent research suggests as many as one in four people in the UK are using platforms like ChatGPT for health guidance, while across the NHS and the private sector clinicians test tools for early assessment, triage, operational planning and workforce optimisation.
He says patient-facing AI is often shallow but broad so far, and that patients want quick, accurate answers at the point of need while remaining in control of their health information.
Webb argues clinically informed design matters, noting generic AI risks spreading misinformation and increasing patient anxiety if left unchecked.
He cites Numan’s work as an example, saying the company built Nu to deliver evidence-based guidance within clinical guardrails and that Numan’s Aegis Monitoring flags risk and brings clinicians in when needed.
Webb adds that AI is valuable for spotting patterns in large datasets to flag early warning signs humans can miss, and that the generative capabilities of AI are already being explored through initiatives such as the MHRA AI Airlock and other validation approaches.
Regulatory Fragmentation And Data Barriers
Webb warns the regulatory landscape lacks end-to-end ownership, with responsibility spread across multiple regulators and oversight functions in England, producing fragmented oversight and unclear expectations.
He says this fragmentation creates inconsistent guidance and slow routes to real-world deployment, which makes it hard for innovators to understand what good compliance looks like in practice.
Webb also highlights data access as a core constraint, saying access to NHS records is inconsistent and private providers are generating clinically relevant data that often cannot be shared back into the wider system.
To address these issues he proposes a more open, human-centred system and suggests an Open Banking-style approach to health data, where consent-led data access would help coordination between NHS services, private providers and patients.
Webb concludes that effective AI governance should pair proportionate regulation with real-world testing, robust monitoring and meaningful human oversight so AI supports rather than substitutes clinical judgement.
