The publication of the NHS’s 10-year plan has put artificial intelligence firmly in the spotlight as one of the Government’s five big bets for transforming healthcare. As a pharmacist working in the UK’s digital healthcare space, this moment has prompted me to reflect on where we truly stand with AI in medicine and healthcare. The parallels with the internet’s adoption journey are striking, yet the pace feels fundamentally different.
The internet took thirty years to move from conception to ubiquity. Each building block had to fall into place: computers needed reasonable processing power, networks required adequate bandwidth, costs had to become affordable, and crucially, people needed familiarity with each piece of technology before mass adoption could occur. We’re witnessing a similar foundation-laying phase with AI in healthcare, but with one critical difference – each technology now accelerates the next.
The New Fundamentals
Where the internet required hardware and connectivity, AI in healthcare demands different fundamentals. Regulations and clinical validation stand as the primary gatekeepers, but perhaps more importantly, we need widespread AI literacy across the healthcare workforce. Not technical expertise, but basic understanding – the equivalent of knowing what electricity is, what it can do, and how dangerous it can be.
Most people can safely plug in appliances and know not to stick their fingers in electrical sockets, yet few could explain how a power grid operates. We need this same practical understanding of AI. Healthcare professionals must grasp what AI truly is without anthropomorphising the technology, recognise its limitations, and retain the ability to question its outputs. We already struggle with questioning authority figures; we cannot afford to fall into the same trap with AI.
The Black Box Problem
The biggest challenge in validating these systems lies in understanding how AI reaches its conclusions. The “black box” nature of many AI models, combined with our professional tendency toward hierarchical thinking, creates a perfect storm. The equivalent of putting your finger in the electrical socket might be feeding live patient identifiable data into general AI models available on the internet, then acting upon the response or allowing AI to alter clinical reasoning without proper safeguards.
Yet we already have frameworks for managing beneficial technologies that carry inherent risks. Medicines, at the right dose, can save lives; at the wrong dose, they become poisons. We accept that statins cause muscle problems in some patients because they prevent heart attacks in many more. Can we develop similar risk-benefit calculations for AI in healthcare?
Professional Responsibility in an Accelerated Timeline
My cautious optimism stems from recognising we’re on the cusp of early adoption of relatively safe AI use, but my concerns centre on human factors. Can regulation and clinical validation move fast enough to prevent patient harm, or will AI acceleration mean they’re always playing catch-up to cutting-edge technology? While I don’t want to impose the equivalent of having a man with a red flag run in front of every car, as a healthcare professional, my first concern must be patient safety.
Professional bodies have a crucial role as bridges between fast-moving technology and slower-moving regulation. The Royal Pharmaceutical Society has recognised this responsibility, recently publishing comprehensive position statements on both “Digital Capabilities for the Pharmacy Workforce” and “Artificial Intelligence in Pharmacy” in early 2025. These documents acknowledge that pharmacy professionals need to understand the opportunities, benefits and risks of artificial intelligence applications for pharmacy practice, whilst emphasising that development of AI tools in pharmacy practice must be co-produced with pharmacists, data scientists, developers, and patients.
The RPS guidance addresses my core concerns about the black box problem directly, stating that it is important that we can critically appraise AI tools, with a focus on explainability, to reduce the risk of the “black box” phenomenon where we are unable to ascertain how the tool reached the output from the original data input. They also warn about AI hallucinations – the incorrect or misleading results that AI models can generate.
Perhaps most importantly, the RPS recognises that our profession must evolve beyond traditional roles. Their digital capabilities framework positions pharmacy professions as more than just a supplier of medicines or a provider of medicines and health information, but as a source of practical wisdom and personal advice. This vision aligns perfectly with managing AI’s integration – we need human wisdom to complement artificial intelligence.
A Vision Takes Shape
This ambitious NHS vision offers an intriguing glimpse of AI’s potential beyond traditional healthcare delivery. The NHS App is positioned as a personal healthcare assistant, supporting and directing patients to appropriate professionals. More fundamentally, the plan envisions using biometric data from wearables to “nudge” the population from the start of ill health towards wellbeing, preventing them from “falling into the river of illness.”
This prevention-focused approach represents a profound shift from reactive to proactive healthcare. Rather than just diagnosing disease, AI could become a constant companion guiding us towards better health decisions before problems arise.
The Primer Awaits
This vision reminds me of Neal Stephenson’s The Diamond Age, where children received a “primer” – an AI technology providing personalized learning tailored to each owner’s needs and development. The primer was powerful yet carefully designed to nurture rather than replace human judgment and growth.
Perhaps this is where we’re heading: AI systems that truly understand individual patients, healthcare professionals, and the complex interplay between them. Not artificial intelligence that replaces human expertise, but augmented intelligence that enhances our ability to heal, prevent illness, and promote wellbeing.
We’re still in the infrastructure-building phase, still developing the professional literacy and regulatory frameworks necessary for safe adoption. But unlike the internet’s pioneers, we’re not starting from zero. The computers exist, the networks are in place, and digital fluency is widespread. The challenge now is ensuring the human elements – understanding, wisdom, and ethical frameworks – keep pace with technological capability.
The primer awaits. The question is whether we’ll build it wisely.
References
Royal Pharmaceutical Society. Digital Capabilities for the Pharmacy Workforce. Available at: https://www.rpharms.com/recognition/all-our-campaigns/policy-a-z/digital-capabilities-for-the-pharmacy-workforce
Royal Pharmaceutical Society. Artificial Intelligence (AI) in Pharmacy. Available at: https://www.rpharms.com/recognition/all-our-campaigns/policy-a-z/ai
LLM’s were used in the creation of imagery and drafting of this blog post.
Leave a Reply