Artificial intelligence (AI) is helping to transform healthcare by improving research, diagnosis, treatment, and care delivery systems. Much of this progress depends on collecting and preparing large health data sets, which underpin the training of algorithms and development of AI models. In this way, health data and AI go hand-in-hand in driving the advancement of next generation treatments and care.
In this report, learn about the new approaches to data management that are helping algorithms train on diverse sets of patient data without exposing private information. Discover how the industry is responding to grassroots efforts calling for algorithmic transparency (empowering people to understand how AI impacts their care regimens) and data equity (ensuring that all communities are fairly represented in the data sets used to train AI systems). And understand how the growing adoption of aggregated data learning—combining different data sets to uncover new insights—and the realization of predictive medicine, are allowing healthcare providers to anticipate health needs for patients and communities before they arise.
Companies are sitting on a wealth of data about their shoppers, their stores and the broader marketplace, but they often lack the internal capabilities and creative insight to put it to use. Successful brands and retailers are differentiating themselves with data-led initiatives like deep-learning algorithms that enable them to quickly adapt or even anticipate shifts in consumer preferences to be first to market with next-gen products inspired by purchase patterns to better serve the needs of their customers, as well as forecast trends.Anticipatory Support
Retailers are applying machine learning and data aggregation tools to anticipate consumer behaviors and needs based on situational or emotional contexts and provide service or product recommendations. This includes using data generated by mobile apps to create a unified view of a member's shopping behavior and preferences, as well as leveraging RFID and other technologies to detect items that shoppers are interacting with in order to target shoppers in the moment and drive incremental retail sales. It also encompasses taking targeted advertising offline with recognition technologies that allow brands to pinpoint and deliver the most relevant ads to specific connected consumers near physical point-of-purchase.Wellness Support
Even before the events of 2020, prioritization of wellbeing was coming to the forefront, for consumers, employers and businesses alike. To meet this need, brands are launching new services around mental health and wellbeing, creating greater points of connection.
In January 2020, Google Health released an AI model, trained on over 90,000 mammogram X-rays, and claimed that it was more accurate in predicting irregular breast cancer presentations than human professionals. In a rebuttal published in the preeminent research journal Nature, over 19 coauthors affiliated with McGill University, the City University of New York (CUNY), Harvard University, Stanford University, and others called for Google Health to share detailed methods and source code behind its AI model, rather than just sharing benchmark results. The rebuttal argued that without model transparency, Google’s research could not be reproduced by fellow scientists and thus verified for its efficacy. The rebuttal marks a key moment in the health industry, calling for algorithmic transparency standards to ensure that AI diagnosis models are safe, consistent, and understood by the professionals employing them.Amazon's app-connected sleep monitor provides personalized insights
Technology leader Amazon is developing a contactless bedside system for tracking sleep apnea. The system uses a device that applies millimeter-wave radar to track people’s breathing and movement. Codenamed "Brahms," the palm-sized tracker displays data on a companion app and even communicates with other IoT devices in the home in the case of a health emergency. More broadly, Brahms can use machine learning to deliver broader insights on a person’s sleep health.By using conversational AI, Watson Assistant for Health Benefits is reducing confusion around health plans and costs
Research shows consumers who don't understand how their health plan works, or how to estimate out-of-pocket costs, are more likely to delay or avoid essential care. Watson Assistant for Health Benefits, developed in partnership with IBM Watson and health insurer Humana, aims to mitigate this issue by leveraging conversational AI to provide Humana members with clear and accurate information on their personal benefits packages, healthcare costs, and available providers—all to help them coordinate timely care. The chat-based service is available to all of Humana's 13 million medical and dental plan members. Live agent support is also available to any member who has trouble with the AI system.