The vast majority of healthcare professionals in the UK are not using AI at work, with 73% reporting that they have never done so – a higher rate than any other nation in the study. This is despite these tools being designed to alleviate some of the biggest issues plaguing the sector – high workloads, burdensome admin and persistent NHS staff shortages.

The report – produced by healthcare AI research and development company Corti and YouGov – suggests that the UK’s reluctance to adopt AI is driven by several factors. A fear of errors is the biggest barrier – with 62% of UK HCPs citing this as their top concern, compared to just 44% in France, 53% in Germany and 56% in Denmark.

Likewise, confidence is also a major hurdle, with less than one in four UK HCPs feeling comfortable with their ability to use AI tools, well below the European average of 31%. Compounding this is a lack of awareness, as 35% of UK HCPs admit they are unaware of how AI could be applied to their work – such as for automating patient notes or supporting decision-making.

The UK’s hesitancy is driven in part by the huge influx of general purpose AI models providing the infrastructure for the majority of apps flooding the healthcare market today – most of which are fragmented tools that do not integrate well with existing systems and are built on general AI models rather than API infrastructures designed specifically for a healthcare setting. As such, many AI solutions are failing to meet the complex needs of over-stretched health workers and ultimately fail to make it out of pilot testing.  

Andreas Cleve, Co-Founder and CEO of Corti, said: "AI is key to the future success of healthcare. In a few short years, we’ve witnessed exciting progress. However, the vast majority of tools available today rely on general-purpose AI that isn’t built to integrate or adapt to the complexities of healthcare. As a result, these tools often fail to deliver, overpromising and struggling to scale beyond trials due to challenges with accuracy, cost, and integration. Is it any wonder clinicians feel they can’t trust AI when the tools they’re given aren’t designed with their needs in mind? Healthcare needs solutions that meet the intricate demands of real-world care environments, built on a new infrastructure the healthcare sector can trust to succeed."

Pressing need: UK staff shortages and burnout

The UK’s hesitation to adopt AI in healthcare is potentially deepening the healthcare crisis – as the NHS continues to grapple with one of the largest staffing shortages in its history and work-related stress costs the organisation £12.1 billion annually.

According to the report, UK burnout rates are among the highest in Europe, at 64% among HCPs over the past year – compared to 41% in Germany and 47% in France. Furthermore, current workloads mean that 26% of UK healthcare professionals think about leaving the field on a weekly basis. Yet the AI tools that could alleviate pressure remain underutilised compared with other European nations.

Germany and Denmark lead the charge

As current European frontrunners in AI adoption, a quarter (25%) of HCPs in Germany and Denmark are already using AI tools at least once a month. A similar study in the US, which has a more open regulatory environment, found that 38 percent of HCPs are using AI in their day-to-day work. In stark contrast, just 14% in the UK use AI once a month.

This gap is likely to be fuelling a growing divide in attitudes, as familiarity and regular usage help to breed trust. HCPs who have used AI tools are more than twice as likely to trust them to support decision making (48% vs 19%) or prompt questions for patients (57% vs 27%), meaning countries like the US, Denmark, and Germany will have head-start advantages when implementing AI at scale in healthcare.

"Just as we wouldn't let a medical student perform surgery without proper training, we shouldn't trust AI systems that haven't been specifically trained and validated for healthcare," said Frederik Brabant, MD and Chief Medical Strategy Officer at Corti. "When over half of healthcare professionals say they wouldn’t feel confident using current AI solutions in their work, we know we need specialised solutions that can build sector confidence and safely meet healthcare's unique demands."

The release of the study coincides with the launch of Corti’s purpose-built AI infrastructure for healthcare. The introduction of three foundation and 20 expert models is the first of its kind, built on nine years of peer-reviewed research, supporting a wide range of administrative tasks – from notetaking and documentation to coding and quality assurance – and factoring in more complex considerations for healthcare that only specialised models can prioritise. Together, Corti’s foundation models form the most powerful and adaptable AI infrastructure in healthcare. Trained on hundreds of millions of hours of domain specific data, the result is faster, more concise, accurate, and cost-efficient AI for healthcare than GPT-4.

A shared vision for AI’s potential

There is evidence that demand is high for AI tools, 74% of HCPs in Europe keen to apply AI to at least one workplace challenge. While top demands among HCPs are for saving admin time (42%) and automating patient notes (29%), many physicians are keen for AI to go beyond just time saving to support decision making and diagnosis (32%) and provide diagnostic insights (25%) in future usage.

Indeed, when asked to imagine AI could reduce their administrative workloads by 80%, 45% of HCPs said they would use the time back to focus on direct patient care.

For more startup news, check out the other articles on the website, and subscribe to the magazine for free. Listen to The Cereal Entrepreneur podcast for more interviews with entrepreneurs and big-hitters in the startup ecosystem.