Legal Knowledge Management and the Rise of Artificial Intelligence
Robotic Refactoring the Workplace
Why Your Next Insurance Claims Processor Could be a Robot
The Forgotten Element in Your Big Data Strategy
"AI -The Future of Automotive Industry"
Nitin Sethi, Global IT Director - Business Transformation & Engagement, Visteon Corporation
WiFi Networks: Shifting from Providing a Service to Improving the...
Daniel J. Strojny, Interim Associate Director of Network and IT Operations, University of St. Thomas
Breaking the Stereotypes in the Development of AI
Yves Jacquier, Executive Director, Production Studio Services, Ubisoft
Operationalize Machine Learning
Zongjie Diao, Director of Product Strategy and Management, Data Center Compute Group, CISCO
Thank you for Subscribing to CIO Applications Weekly Brief
Some problems can be solved best using a set of established rules, which can be categorized as a deterministic approach. Rest of the problems need analytical or probabilistic approaches that can get more accurate as additional data is fed into the system.
FREMONT, CA: With the advancement of the technology, people witness blockchain, deep learning, natural language processing, computer vision, and a slew of other artistic terms that AI is bandied to as an answer to the healthcare challenges. But these technologies are simply the tools to solve some part of the problems, and are not always the best ones to sort a specific problem completely.
Complicated tasks such as automating the work of medical scribes or patient coordinators need sophisticated AI techniques like ML, natural language processing, and language comprehension. In the same manner, helping a radiologist to make precise readings of the volume of images generated by the latest imaging solutions can make image processing and ML applicable. Probabilistic algorithms can be trained mechanically with new data to adapt, whereas the deterministic rules need careful manual modification by experts.
Chatbots can be used to reschedule an appointment or have interaction regarding insurance authorization with only the patient. In such a case, probabilistic approaches that get it wrong occasionally might be acceptable. But if a chatbot is left to automate a clinical conversation, then the poor accuracy can lead to severe consequences. In such situations, human oversight is a significant factor to review the conclusions by the chatbot, unless the machine is entirely mechanized to generate accurate results.
Predictive and probabilistic approaches in healthcare provide valuable and actionable insights that can be adapted to engage patients in their healing journeys. But extreme care is necessary for the clinical scenarios that involve diagnosis or treatment planning because an algorithm that assists a radiologist is much different from the one that intends to replace it.
The most overlooked aspect while using an AI-driven method is user experience, especially with the probabilistic black-box approach such as ML. When an ML algorithm suggests a diagnosis with no explanation on how it arrived at that recommendation, it is natural for the provider and patient to hesitant in accepting that guidance. Whereas, deterministic rules-based solutions can justify the reason behind the suggestion with the help of their most consistent and explicit set of logical statements.
AI-driven healthcare can undoubtedly get more robust over time and improve the effectiveness of the surrounding in the broader ecosystem as well as enhance the outcomes for patients, which is the ultimatum.