Home / Health / AI Scribes Risk Patient Safety With Errors
AI Scribes Risk Patient Safety With Errors
31 Aug
Summary
- AI scribes mistakenly record drug and illness errors.
- Patients, not GPs, often identify AI transcription mistakes.
- AI's accuracy is questioned, risking patient medical records.
AI technology designed to transcribe patient-doctor consultations is raising significant patient safety concerns. Healthwatch England has highlighted instances where AI scribes have made critical errors, such as misidentifying drug names and medical conditions, with patients often being the first to notice.
One patient experienced significant distress when an AI scribe wrongly suggested a demyelination diagnosis, a serious nerve condition. This error, and others like it, could inadvertently enter patient medical records and affect future care if not detected and corrected.
Despite the government's plan to integrate AI scribes for efficiency, concerns persist about their accuracy. The UK is using 27 different AI scribes, and ministers have been warned of potential legal repercussions for the NHS and medical professionals due to AI-related mistakes.
Experts acknowledge that AI can make mistakes, particularly with complex cases or non-native English speakers. However, the need for thorough human review means AI tools are not yet proving to be time-saving, according to some GPs.
