“At breaking point”, “collapsing”, “medieval inequalities”. These are the headlines dominating Britain’s front pages when it comes to the NHS. The health service that once defined post-war progress now finds itself battling staff burnout, funding pressures, and a digital transition that is far from smooth. How did it get here?

A study attributed these challenges to “years of poor workforce planning, weak policies, and fragmented responsibilities.” The scale of the crisis is underscored by the wave of industrial action. This month, NHS resident doctors are staging a five-day strike demanding fair pay, improved working conditions, and more stable career prospects. The British Medical Association has warned of a “looming unemployment crisis” for UK physicians, noting that last year saw 30,000 applicants fight for just 10,000 training posts.

Despite a decline in the number of general practitioners, the NHS continues to face mounting backlogs, and the number of patients needing care has not fallen proportionally. Reports showed that in 2022, the NHS delivered more primary-care appointments than before the pandemic, placing even greater pressure on an already overstretched workforce. Unsurprisingly, many doctors, exhausted and increasingly disillusioned, are now choosing to move abroad. A recent British Medical Association survey found that 30% of doctors are very or fairly likely to relocate overseas within the next year, and a further 33% said they are considering doing so at some point in the future. As more clinicians contemplate leaving, staffing shortages intensify, waiting lists grow, and the system’s overall resilience continues to erode.

Technology as the Proposed Solution

In the face of these pressures, one solution repeatedly proposed is the adoption of new technologies, particularly AI. The hope is that smarter systems can ease administrative burdens, optimize patient care, and enhance diagnostic accuracy, whilst alleviating some tasks from overburdened physicians.

However, while AI holds immense promise, its rollout has repeatedly collided with a stubborn reality: “the NHS still struggles with basic barriers”, explains Dr Chloe Jacklin, a former NHS physician now working in the healthtech industry. Wi-Fi patchiness, old programs and computers, as well as non-transferrable systems access, are some of these issues. Dr Jacklin continues to work locum shifts within a London NHS Trust, giving her a front-row view of the gap between aspiration and reality.

“Right now, I am not doing direct patient work full-time because I feel that the healthtech industry is making more progress in improving patient care than what I am seeing elsewhere. The industry makes me feel optimistic.”

When the Basics Don’t Work

The problem starts with something as mundane as connectivity. A 2025 survey found that UK healthcare workers lose over 122 hours per year, which is about 17 working days, due to outdated and unreliable digital infrastructure. More than half of those surveyed said these delays affected their ability to provide timely care.  

At this year’s BMJ Future Health event, Dr Ally Jaffe, an NHS psychiatry resident doctor, recalls just how bad it can get:

“I remember being on call, and there was a patch of hospital that had no Wi-Fi signal. So we used walkie-talkies to communicate with staff. I remember thinking: 'What age am I living in?'"

Logging on isn’t the end of the story; Dr Jacklin points out that old computer programs slow things down even more. In an interview with us, she says:

“For instance, the intensive care unit uses a computer program that was designed by a doctor in the early 2000s. It’s really clunky, and not very useful; it crashes a lot.”

Connectivity issues run deep. Dr Jaffe further explains a problem that happens before you even step inside the hospital:

“If I'm in North London at Highgate and I go to University College Hospital, there's a completely different interface system. Even from a basic level, it requires a different key card. So, I can't even get to the ward I need to."

She would have to wait for another member of staff to let her in, which “significantly stumps” productivity. Once she finally makes it to the ward, she doesn’t have the login details for the system. “More waiting,” she sighs.  

When the basics - such as network access, hardware reliability, security permissions - fail, it’s hard to see how AI could be incorporated within this fragmented framework.

It’s a Two-Way Street

If the digital infrastructure were better, would we see improvements? Maybe not. In fact, a lack of technological understanding could exacerbate the situation.

A 2023 study published in the National Bureau of Economic Research tested this. Researchers gave radiologists access to an AI tool that outperformed most human readers in image interpretation. Yet, paradoxically, when radiologists assisted the AI in this task, they did not achieve higher overall diagnostic accuracy, and their decision-making took significantly longer. In other words, access to AI did not translate to better performance, underscoring that technology must be accompanied by training, workflow design, and trust to deliver genuine value.

The results of this nod to the broader issue of digital literacy in the NHS, or the lack thereof. A recent OECD paper highlights that staff should be equipped to adapt to digital changes as healthcare roles evolve. However, according to an article published by Health Tech World, NHS physicians and nurses are, on average, less digitally literate than the wider UK workforce. Without foundational digital skills, the NHS will struggle to take up new tech.

Cyber Risks

Simultaneously, digital illiteracy poses a cyber-safety threat. The casual use of generative AI tools in clinics is a growing concern as doctors, seeking efficiency, may lean on these systems without fully understanding privacy implications.

As widely used as it is, using an LLM, like ChatGPT, comes with its own risks. Inputting any patient data, even when identifiers like names are removed, can still violate data-protection regulations, since the information passes through external servers beyond the NHS’s control. As Jacklin warns:

“Physicians shouldn’t be putting patient information into ChatGPT. Some people do it because it is obviously quicker and more convenient, but even if you take out patients’ names, it is still not fully secure.”

Cyber issues extend to the systems themselves, according to Jacklin. She notes these “legacy programs” come with “minimal protections,” which make them susceptible to data leaks.

In 2020, the first known “cyber-murder” was investigated in Germany after a ransomware attack paralyzed the computer systems at Düsseldorf University Hospital. The attack blocked access to critical systems and patient records, forcing the hospital to divert incoming patients to other facilities. One critical ill patient had to be transferred 30 kilometers away, and the delay proved fatal for one critical patient. The German authorities opened a manslaughter investigation, underscoring the stark reality that cybersecurity failures in healthcare are not just technical incidents; they can cost lives.

It is the hospitals and the NHS’s responsibility to put tighter regulations regarding cybersecurity and create a guide for physicians. As the systems and their users are already vulnerable to cyber threats, there needs to be a safety net in place. Those who work in an office usually go to their trusted IT department in these instances; the process, more often than not, is seamless.

Healthcare workers are not so lucky. “Attracting IT professionals to the NHS can be challenging, and hospital IT teams are often overwhelmed.” Dr. Jacklin notes.

The problem then bleeds into the problem of AI adoption itself: dedicated IT teams are usually the first to advocate for new technology and fill a mentor role as the rest of the organization gets to grips with unfamiliar digital tools. But without a dedicated, adequately funded technical team, this doesn’t happen.

“If hospital staff are too busy fighting fires and trying to deal with problems that are already going on, they don’t have much headspace to think about the next steps, to be proactive, thinking about the future,” says Jacklin.  

Digital Disarray and Workforce Hesitation

Technical frustration due to outdated digital infrastructure feeds into deeper cultural and psychological barriers. Clinicians, already stretched thin, are understandably wary of another digital system that could slow them down. NHS England acknowledges this tension:

“Staff may be reluctant to adopt AI technologies if they feel threatened or unconvinced, if they are worried about the risks, or if they do not see enough evidence of effectiveness. They need to be brought on board so that those who are worried feel empowered to shape how the technology can be used to support them.”

Trust is earned through experience, and frontline workers need to feel assured and to have the burden alleviated from them of the question of trust in emerging technologies.  

Ultimately, this brings the wider theme into focus: the doctor-machine relationship. Even with the right infrastructure and clear governance, how clinicians perceive, trust, and interact with AI is what determines whether these technologies enhance care or simply add friction.

The Way Forward

So, what can be done?

Dr Jacklin argues that meaningful progress requires action on several fronts and suggests three pragmatic steps. First, digital transformation needs to be driven from the top, with stronger political leadership, clearer accountability, and nationwide endorsement of safe technologies.  Second, she emphasizes the need for investment in infrastructure: replacing outdated computers, strengthening Wi-Fi across hospital estates that were, as she notes, “built before wireless connectivity even existed.” Lastly, she stresses the importance of targeted training and governance, in which cybersecurity and digital literacy education should be embedded into mandatory training for clinicians.

The NHS’s 10-Year Plan does recognize digital transformation as a central goal, yet critics argue it lacks detail on how these ambitions will be achieved. Without adequate funding and coordination, plans risk remaining aspirational.

Conclusion

The NHS does not suffer from a lack of innovation or enthusiasm for technology. It suffers from a lack of readiness. Reliable infrastructure, interoperable systems, and a digitally confident workforce are prerequisites for AI to deliver its promise. Until the basics are addressed, every new system risks becoming another burden.  

AI may one day transform British healthcare. But first, the NHS must ensure that its digital foundations are strong enough to hold it. Otherwise, trying to modernize will remain what it often feels like today: like icing a hot cake, it won’t stick.