Dutch Doctors Turn to Artificial Intelligence to Record Patient Consultations

2026-08-14 digital

Utrecht, Friday 14 August 2026
A Nivel survey reveals 36% of Dutch GP practices now use AI voice-to-text tools, saving doctors nearly 43 seconds per consultation amid unresolved privacy regulations.

The Rapid Rise of AI in General Practice

A significant shift in the Dutch primary healthcare sector has been highlighted by the annual Nivel General Practice Survey, published on 13 August 2026 [3]. The study, which surveyed over 800 of the approximately 4,500 general practitioner (GP) practices across the Netherlands, revealed that 36 per cent of clinics now utilise AI-driven voice-controlled reporting tools to transcribe patient consultations [1][2]. This represents a substantial surge from the 14 per cent adoption rate recorded in 2024, representing an increase of 157.143 per cent in just over a year [2][4]. Overall, 46 per cent of Dutch GP practices now use at least one AI-supported digital application, up from approximately 20 per cent in 2024, which is an increase of 130 per cent [4][5]. While voice-to-text documentation has led this technological wave, AI-supported triage has also risen to 16 per cent from 10 per cent over the same period [4].

Saving Time and Reducing Cognitive Load

For overworked healthcare professionals, the primary appeal of integrating these advanced SaaS tools into their clinical workflows is administrative relief. A landmark study conducted by Reinier van Linschoten at Erasmus MC, which analysed 535 consultations, demonstrated that AI note-takers save doctors an average of 42.7 seconds per consultation by eliminating manual typing [1][2]. Van Linschoten, a university lecturer and GP-in-training, noted that while this technology does not necessarily enable doctors to see more patients, it provides vital ‘breathing room’ and reduces cognitive load [1][2]. Furthermore, Frank van Harmelen, an AI professor at Vrije Universiteit, pointed out that the technology enhances the social dynamic of consultations, allowing doctors to maintain direct eye contact with patients rather than focusing on a keyboard [2].

Clinical Limitations and the Danger of Complacency

Despite the clear efficiency gains, the technology is not without its limitations. The Erasmus MC study revealed that while AI models are highly comprehensive at capturing patient histories, diagnoses, and action plans, they are significantly less effective at documenting physical examinations, such as blood pressure readings or wound inspections, unless the physician explicitly narrates these actions aloud during the consultation [2][6]. There are also concerns regarding software reliability and the risk of ‘hallucinations’—where the AI generates non-factual information—although no dangerous medical errors have been reported to date [1][6]. Professor van Harmelen warned of human complacency, noting that if the previous 300 draft summaries generated by the AI were accurate, a GP might fail to critically review the 301st entry [2]. The Health and Youth Care Inspectorate (IGJ) confirmed that as of August 2026, it had received zero incident reports regarding AI voice-controlled reporting tools [1][2].

Regulatory Hurdles and Privacy Debates

The rapid digitisation of patient records has triggered intense debate surrounding data privacy and medical confidentiality under the General Data Protection Regulation (GDR). Digizo, a Dutch digital healthcare knowledge platform, has formally approved several AI voice tools, arguing that privacy risks are low because raw audio files are not stored and reports must be verified by a GP before being saved [1][2]. Consequently, Digizo suggests that explicit patient consent may not be strictly mandatory [1][6]. However, the Dutch Data Protection Authority (AP) has refused to endorse this stance [2][6]. An AP spokesperson emphasised that doctors are legally barred from sharing patient data with third parties—such as external cloud hosting providers—without explicit consent, and confirmed that the regulator is currently drafting comprehensive guidelines to address AI implementation in the medical sector [1][2].

Sources & Ecosystem Partners

Digital Health Clinical Workflow