For decades, speaking the Scots language, along with its rich regional variants like Doric, in a formal medical setting has come with a hidden, subtle tax. When a patient in the North East tells a doctor they have a “gey sair wame” or describes a dizzy spell with local idioms, the nuance can be lost if the clinician is from outside the area or relying on standard medical transcription software. Language barriers do not just exist across borders; they exist across postcodes, and in healthcare, being misunderstood can lead to genuine inequalities in patient outcomes.

This month, a trial at Inverurie Hospital’s Donbank Ward has introduced a radical solution: an artificial intelligence system that is actively learning to understand Doric. In a cultural climate where AI is rightly viewed with deep suspicion by Scottish creatives and language advocates, it is vital to understand exactly what this technology is doing and why its pedigree across Europe proves it is a medical lifeline.

The European Blueprint: From Copenhagen to the Alps

To understand why the Doric trial is so promising, you must look at the technology’s track record across the broader Germanic language sphere. The Scots language, much like Danish and German, is a distinct Germanic language characterised by complex regional variations, shifting vowel sounds, and unique syntax that standard AI models traditionally fail to grasp.
The system powering the Inverurie trial, developed by health technology firm Corti, has already conquered some of the most difficult linguistic environments in Europe:

Denmark’s 112 Success: In Copenhagen’s emergency medical dispatch centres, the AI listens to live 112 calls. Navigating regional Danish variants and the chaotic background noise of emergencies, the AI proved capable of detecting out of hospital cardiac arrests with up to 95% accuracy. This is significantly faster and more accurately than human operators relying on standard protocols.

The Swiss German Challenge: Recently, the technology was rolled out across Switzerland in partnership with Voicepoint to support thousands of clinicians. Swiss German is notoriously fragmented, with local tongues changing dramatically from valley to valley. The AI successfully bridges the gap between spoken Swiss German and written Swiss High German medical records, reducing documentation time by up to 80%.

German Clinical Integration: Across Germany, the AI is deeply integrated into practice management systems, turning regional German speech directly into structured, compliant medical letters.
If the technology can untangle the linguistic web of Swiss valleys to assist doctors in the Alps, its application in Aberdeenshire is built on a foundation of proven, medical grade utility.

Listening, Not Replacing
It is easy to hear the words “AI” and “Scots language” and immediately picture the technology scraping the internet to generate automated poetry, displacing local writers. Those are real, valid threats to the creative industry that demand robust legislation.
However, the application in NHS Grampian is fundamentally different. This is not a generative engine trying to create culture; it is a highly advanced dictation tool trying to capture clinical facts.

It gives time back: Nurses at Donbank Ward report that documentation time has dropped by nearly 70%.

It requires human oversight: The AI does not finalise the notes. It drafts them, and the nurse reviews, edits, and corrects them. Every time a nurse corrects the system to properly understand a Doric phrase, the system’s local intelligence improves.

It does not make decisions: The technology cannot diagnose or prescribe. Its sole purpose is to handle the administrative burden.

Rehumanising the Hospital Ward

The most profound feedback from the Inverurie trial has not been about technological processing power, but about eye contact.

When healthcare professionals are freed from the necessity of staring at a screen or a clipboard to jot down notes, they can look at their patients. They can hold a hand, read body language, and engage in meaningful conversations focused on the individual. For elderly patients or those in distress, having a nurse who is fully present in the moment is invaluable.

If a machine can take on the robotic task of data entry so that a nurse can be more human, it is a victory for patient care.

A Step Towards Language Equality

Recognising the validity of the Scots language means recognising that speakers deserve to use their mither tongue when they are at their most vulnerable, such as in a hospital bed, without fear of being misheard or causing an administrative headache.
The broader debate about AI and copyright will rage on, and the creative community must continue to fight for the protection of their work. But within the walls of Inverurie Hospital, teaching a machine to understand Doric is a monumental step forward. It ensures that every patient, whether they proudly consider themselves a “teuchter” or simply speak the local tongue, receives the exact same standard of care, safety, and attention.