BE PART OF THE REVOLUTION TRANSFORMING HEALTHCARE WITH AI - PowerPoint PPT Presentation

BE PART OF THE REVOLUTION TRANSFORMING HEALTHCARE WITH AI CALIFORNIA — THE RITZ-CARLTON, LAGUNA NIGUEL 11–14 DECEMBER 2019 1000 ATTENDEES 80 SPEAKERS 10 WORKSHOPS www.aimed.events/northamerica-2019/ 2 SOCIAL EVENTS #AIMed19 1 AIMed19

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019 CIO/CxO Workshop AI Implications in Healthcare John Henderson, Vice President & CIO CHOC Children’s Company Name www.aimed.events/northamerica-2019/

AIMed NORTH AMERICA, CALIFORNIA 11–14 DECEMBER 2019 Is Is Data ta In Inte tegrity ity Imp Importa tant? t? 70% What percentage of medical records have errors? • Poor quality input may produce unexpected or erroneous AI output • www.aimed.events/northamerica-2019/

Ap Apply lyin ing AI AI & Machin hine e Lea Learnin ing to o Clin linic ical l Ca Care & O & Ope perations s Ap Applications of f AI AI in Healthcare Pre-authorizations • Payments • Coding efficiencies • Readmission Risk • Early sepsis predictor • Drug diversion •

AI AI Use se Cases ses in Hea Healthcare AI Use Cases fo AI for Healthcare • AI-Enabled Diagnostic Imaging Interpretation • Deep Learning Algorithms • Virtual Personal Health Assistants • Augmented reality, cognitive computing, sentiment analysis, and speech and body recognition to create a virtual encounter • AI for Virtual Care Monitoring and Real-Time Operations Management • Predictive and prescription alerting drive decisions for improved outcomes

Th The Ethics cs of f Ar Artifi fici cial Intelligence ce The C Th Case se f for Eth Ethics i s in A n AI Key Considerations: Bias can relate to race, gender, age, location or time. • It can also favor a specific data structure or a specific problem to solve. • Bias is a natural effect of learning. Eliminating it is no possible. • Approach it by using the bias and target creation of multiple algorithms • to adjust for the bias you want to avoid Trustworthy results come from a diversity of algorithms working on the • same problem Some forms of bias may be desirable — for instance, when • determining your values as part of the learning process i.e., wanting to avoid bad language and favoring empathic, polite and patient language are forms of bias toward what you rightly think is important for a conversation or content you want delivered or in moderated conversations between AI-enabled systems and people Data for AI often contains incomplete or biased information because • data sources are insufficiently diverse.