The question. Can machine learning improve the diagnosis of laryngeal disorders from endoscopic and stroboscopic images, and what are the ethical limits of using generative AI in the field?
The approach. Development and assessment of deep-learning models for laryngeal diagnosis, beginning with sulcus vocalis on videostroboscopy, together with scholarship on the responsible use and dissemination of generative AI in otolaryngology, and studies of the quality of AI-generated patient information.
Current status. Published work includes Artificial intelligence based diagnosis of sulcus (European Archives of Oto-Rhino-Laryngology, 2024), Using Generative Artificial Intelligence in the Production and Dissemination of Innovation in Otolaryngology, Ethical Considerations (Otolaryngology–Head and Neck Surgery, 2024), and studies assessing large language models as sources of patient information. See Publications for links.
Collaboration. We are interested in collaboration on datasets and external validation of laryngeal diagnostic models.