2020 Responsible Machine Learning Summit: AI and COVID-19
John O'Donovan, Senior Director, AI at Teladoc Health
AI and Telehealth: State of the Art in Provider Augmentation
Abstract: Covid-19 has changed the future of healthcare in a dramatic way, rapidly accelerating the inevitable shift from traditional brick-and-mortar hospital visits, to simple, accessible interactions between patients and providers over smartphones and other devices. This shift in the provision of care enables us to empower providers with AI driven insights that are typically not accessible in traditional consults. By creating a layer of multimodal AI services between the doctor and patient, we can assess medically-relevant signals that may otherwise go unnoticed. For example, doctors can not see respiration, pulse or temperature, but AI-driven sensors can accurately read these from a smartphone camera and provide the information to the provider in real time. In the case of stroke, a critical indicator is the advance of facial asymmetry over time. This can be difficult to detect, even for a keen-eyed neurologist. AI models can apply facial tracking to automatically monitor for changes in asymmetry over time, helping to alert medical staff in acute situations, improve time-to-treatment and potentially save a life. At Teladoc, we host over 10 million consults per year, some in hospitals and some from home. Using state of the art speech and natural language models over this huge and highly unique data resource, we have the potential to ensure the highest standards of clinical quality at scale, provide insights into critical actions such as opioid and antibiotic prescription rates, and train a variety of classifiers to improve the standard of care on our platform in many new and exciting ways.
Biography: John is a senior director and technical lead for AI at Teladoc, which is currently an US $18B company and expected to double that later this year in a large merger with Livongo inc. John is also a co-founder of Siras Systems, which is a Goleta-based company that provides personalized educational planning for children with special needs. He holds an appointment in the Computer Science department at UCSB as an Associate Researcher, where he previously co-directed the Four Eyes lab with Professors Turk and Hollerer. He received his PhD in 2008 with a focus on Trust in Recommender systems, and has published over 100 peer reviewed papers. He was program co-chair of ACM RecSys in 2018, IntRS 2020 and ACM IUI in 2021 and serves on the steering committee for these. His research has received 6 international awards for best or outstanding papers the last 5 years.