Content area
Abstract
[...]these notes are increasingly burdensome to write, thanks in part to the electronic health record (EHR). AI-generated notes are not transcriptions; like clinicians, the scribes group distinct problems and split appointments into recognisable components, creating the visit narrative. [...]in addition to reducing administrative work, some argue AI scribes can create more time for the clinical attention. Generative AI scribes do not create narratives from scratch; they incorporate information from current recordings, past examples of medical notes, preference and ranking data provided by human annotators, large sets of internet training data, and patterns encoded in the AI model. Most immediately, seasoned clinicians will need to reduce transcription errors, odd word choices, extraneous details, and disclosures inappropriate for the record as well as adding missed details and often rewriting AI-generated assessment and plan sections, which contain the cognitive work of differential diagnosis building, test selection, plan formulation, and patient education.
Details
1 English Department and Health, Humanities, and Society Program, Northeastern University, Boston, MA 02115, USA
2 Codman Square Health Center, Dorchester, MA, USA
3 Pioneer Centre for AI, University of Copenhagen, Copenhagen, Denmark
4 Department of Quantitative Theory and Methods and Department of English, Emory University, Atlanta, GA, USA