COVID-19 has illuminated the need for clearer AI-based risk management strategies

Abstract

Machine learning methods offer opportunities improve pandemic response and risk management by supplementing mechanistic modeling approaches to pandemic planning and response based on diverse sources of data at every level from the local to global scale. However, such solutions rely on the availability of appropriate data as well as communication and dissemination of that data to develop tools and guidance for decision making. A lack of consistency in the reporting and availability of disaggregated, detailed data on COVID-19 in the US has limited the application of artificial intelligence methods and the effectiveness of those methods for projecting the spread and subsequent impacts of this disease in communities. These limitations are missed opportunities for AI methods to make a positive contribution, and they introduce the possibility of inappropriate use of AI methods when not acknowledged. Going forward, governing bodies should develop data collection and sharing standards in collaboration with AI researchers and industry experts to facilitate preparedness for pandemics and other disasters in the future.

Keywords:

• COVID-19
• Artificial Intelligence (AI)
• AI applications
• disaster response

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This material is based upon work supported by the National Science Foundation Graduate Research Fellowship under Grant No DGE 1841052. Jon Zelner was supported by grant No. 6 U01IP00113801-01 from the National Center for Immunization and Respiratory Diseases at the US Centers for Disease Control and Prevention. This work is also supported by a COVID Skunkworks grant from the University of Michigan College of Engineering. This support is gratefully acknowledged. However, views and opinions in this paper are those of the authors and do not necessarily reflect the views of the sponsor.