Artificial intelligence (AI) in healthcare is rarely out of the headlines. From AI-assisted clinical documentation to intelligent patient triage, the conversation often focuses on what the technology can do next.
As AI capabilities continue to accelerate, new research from the Medicines and Healthcare products Regulatory Agency (MHRA) looking at public views of AI in healthcare suggests the real challenge is no longer technical. It’s earning the confidence of the people expected to use it.
Rather than asking whether patients are ready for AI, the research suggests they already recognise its potential. Instead, their biggest concern is whether they can trust it.
The MHRA’s extensive public engagement programme, which gathered views from patients, clinicians, industry and healthcare organisations across the UK, found broad support for AI in healthcare, provided it is introduced safely, transparently and with appropriate human oversight.
That evolves the conversation. Success won’t be defined by adopting the latest AI tools or deploying them the fastest. It will depend on how confidently patients, clinicians and practice teams can trust the systems surrounding them.
What does the public think about AI in healthcare?
The short answer is that people see genuine value in AI, as long as it improves care without compromising safety.
The MHRA research found strong recognition that AI in healthcare could help improve quality, speed and convenience of care. Participants also acknowledged its potential to reduce pressure on NHS staff and support better patient experiences.
This reflects the reality many patients already experience today. Technology is already embedded across healthcare, whether that’s booking appointments online, navigating practice phone systems, receiving appointment reminders or accessing health information digitally.
Patients aren’t deciding whether to trust technology for the first time. They’ve already learnt to trust digital services when they are reliable, transparent and backed by clear accountability. AI in healthcare is simply the next step in that journey.
Patients are already turning to AI for health advice, highlighting growing expectations for faster, more accessible support. The challenge isn’t convincing patients that AI belongs in healthcare. It’s ensuring it improves care in ways they can understand, question and trust.
Why trust matters more than technology
One message stood out consistently throughout the MHRA findings: trust is essential. But trust isn’t confidence that AI will never make mistakes. It’s confidence in the safeguards surrounding it.
Patients and clinicians want reassurance that AI in healthcare is used responsibly. That means being clear when AI is involved, maintaining meaningful human oversight, protecting patient information and ensuring technologies continue to be monitored after they are introduced into real clinical settings.
Importantly, the public wasn’t calling for innovation to slow down. Instead, they wanted regulation to keep pace with technological change, balancing innovation with robust safeguards.
This reflects a broader shift in healthcare technology.
It’s no longer enough for an AI solution to demonstrate impressive capabilities. Healthcare organisations increasingly need confidence that products have been developed, assessed and deployed against recognised clinical safety, governance and regulatory standards. The MHRA’s research reinforces this point. Building safe AI in primary care depends on transparency, clinician oversight and appropriate governance throughout a product’s lifecycle.
We’re already seeing this across the industry. Recently, the TORTUS technology powering Surgery Intellect became the first ambient voice technology (AVT) to achieve UKCA Class IIa certification. While clinicians remain responsible for reviewing and approving every AI-generated note before it is added to the patient record, independent assessment provides additional assurance that the technology has been evaluated against the UK’s medical device framework.
Trust also depends on the safeguards built into AI itself. Surgery Intellect is powered by TORTUS which includes The Shell – a clinical safety layer designed to identify unsupported AI outputs before they reach the clinician. Combined with clinician review and independent assurance, these layers of oversight help ensure AI supports clinical practice safely and responsibly.
Ultimately, trust isn’t something organisations can add later. It has to be designed into technology from the beginning.
How can AI in healthcare improve patient access?
While headlines often focus on AI replacing jobs, the greatest opportunity may be far less dramatic.
Much of healthcare’s administrative workload doesn’t require clinical judgement. It involves repetitive, manual processes that prevent clinicians and practice teams from spending time where they’re needed most.
This is where AI can make a meaningful difference.
Across primary care, AI is already helping practices:
- Navigate patients to the most appropriate service
- Reduce unnecessary telephone queues
- Summarise conversations into structured documentation
- Support more consistent triage workflows
- Identify patterns in demand to improve workforce planning
- Automate routine administrative tasks
The goal isn’t to replace people. It’s to remove unnecessary friction from the patient journey, and to streamline repetitive administrative processes.
For patients, that means spending less time waiting in queues, repeating information or navigating complex access routes. For practice teams, it means having more time to focus on the conversations and decisions that require human expertise.
This aligns closely with wider NHS ambitions to use AI in healthcare where it delivers practical benefits while maintaining clinical oversight and local accountability. NHS England’s guidance on Ambient Voice Technology1, for example, emphasises safe implementation through clinical governance, integration with existing workflows, and clinician review rather than autonomous decision-making.
As these technologies mature, organisations are likely to be judged less on whether they use AI and more on how confidently they can demonstrate that it is being deployed safely, responsibly and with appropriate oversight.
Will AI in healthcare replace the professionals?
No, AI in healthcare won’t replace professionals. If anything, the MHRA findings reinforce the opposite.
Throughout the research, participants consistently supported AI assisting healthcare professionals rather than replacing them. Human judgement, accountability and oversight remained fundamental expectations.
We believe this reflects the future of primary care: AI-assisted, human-led.
Technology should reduce repetitive work, surface useful information and streamline workflows. Clinicians should remain firmly in control of diagnosis, treatment decisions and patient care.
The same principle applies across patient access. Intelligent technology can help patients reach the right service more quickly, but the responsibility for clinical decisions should always remain with appropriately trained healthcare professionals.
When AI supports people instead of replacing them, everyone benefits.
Building trustworthy AI in healthcare
The MHRA’s research changes the conversation around AI in healthcare.
Patients are not rejecting innovation. They are asking healthcare providers to implement it responsibly.
That responsibility extends beyond choosing the most capable AI tools. It means choosing technologies that are transparent about how they work, designed with patient safety in mind, supported by appropriate governance and continuously monitored as they evolve.
For organisations adopting AI, trust should never be treated as a marketing message. It should be built into every stage of the product lifecycle, from design and regulatory assurance through to implementation and ongoing monitoring.
The NHS faces growing demand, workforce pressures and increasing expectations around access to care. AI has an important role to play in addressing these challenges, but its long-term success won’t be determined by model performance alone.
It will depend on whether healthcare organisations can show patients and clinicians not just what AI can do, but how they have ensured it can be used safely, responsibly and with confidence.
The future of AI in healthcare won’t belong to the organisations using the most technology.
It will belong to those that earn the greatest trust.
Not all healthcare AI is regulated in the same way. Explore our guide to medical device classifications and what they mean for your practice.