{"id":114300,"date":"2020-06-17T13:19:18","date_gmt":"2020-06-17T13:19:18","guid":{"rendered":"https:\/\/news.microsoft.com\/?p=437950"},"modified":"2020-06-17T13:19:18","modified_gmt":"2020-06-17T13:19:18","slug":"using-ai-to-help-governments-think-through-covid-19s-impact-and-strategies-for-recovery","status":"publish","type":"post","link":"https:\/\/sickgaming.net\/blog\/2020\/06\/17\/using-ai-to-help-governments-think-through-covid-19s-impact-and-strategies-for-recovery\/","title":{"rendered":"Using AI to help governments think through COVID-19\u2019s impact and strategies for recovery"},"content":{"rendered":"

In any time of uncertainty, it is critical we leverage data-driven approaches when solving problems. About a decade ago, we wrote on this blog<\/a> how we used a data- and model-driven approach to guide us to the cloud as the future of enterprise computing. Today, we\u2019re applying the same foundational approach while benefiting from the power of our cloud and AI capabilities to unpack today\u2019s source of great uncertainty: COVID-19. In the white paper we are releasing today<\/a>, we outline a policy framework to help governments think through the impact of COVID-19 and recovery strategies. We have also included an economic model that quantifies economic impact.<\/p>\n

Economic impact framework<\/strong><\/p>\n

As the white paper highlights, our economic impact model is built on a large dataset of economic signals and takes insights from economics and epidemiologists to define scenarios and estimate GDP impact. We ingest real-time data as well as traditional economic data, use our Azure AI capabilities to drive insights, our Azure cloud to run calculations, and we update the latest numbers weekly on our Power BI dashboard. We continue to evolve and improve the model as we ingest new signals and learn more from others through our discussions with policymakers and NGOs. But policymakers don\u2019t merely predict GDP; they help shape it. As we engaged with them, we saw the need for a framework and data that helped them navigate this.<\/p>\n

\"covid<\/a><\/p>\n

Social contact budget<\/strong><\/p>\n

\u201cSocial contact\u201d used to be something we didn\u2019t have to think about. It was a byproduct of going to the store or the gym, often viewed as a positive byproduct. Since the start of COVID-19, it comes at a high price. Today, it can perhaps be compared to carbon emissions: an unwelcome byproduct of economic activity. One can lead to pollution; the other to infection. In economic terms, social contact has become a scarce resource. It has become the linchpin between managing infections and protecting the economy \u2013 it is what is driving up infection rates but is also needed for economic activity. By treating it as an economically scarce resource, it raises three critical questions that we began to address in the paper in a data-driven way:<\/p>\n

    \n
  1. How much room do we have to open the economy (\u201csocial contact budget\u201d)?<\/li>\n
  2. How do we best spend the budget (\u201creturn on social contact\u201d)?<\/li>\n
  3. How do we grow our social contact budget over time (\u201creducing cost of social contact\u201d)?<\/li>\n<\/ol>\n

    On the first question, the lower the transmission rate \u201cR\u201d (R being the average number of secondary cases per infectious case), the greater the social contact budget<\/em> and thus the more room there is to open up parts of the economy while avoiding a second wave of infections, which is very costly from a health and an economic perspective. Second, as is true with any budget, we must spend it wisely. Depending on what a policymaker optimizes for (e.g. GDP, employment, avoiding bankruptcy), we created data-driven views on how to optimize return on social contact<\/em>. Industries that can work from home should work from home, as the ones who cannot need the social contact budget more. For the ones who really depend on social contact, we should use data to inform decisions. For example, the figures below compare various industries on their propensity to drive social contact vs. GDP.<\/p>\n

    \"cost<\/a><\/p>\n

    Finally, over time, the budget can be expanded through changes in behavior as people adjust to the \u201ccost of social contact\u201d and through measures such as massive testing and contact tracing. These measures essentially weaken the relationship between social contact and R. They ensure that social contact happens between healthy people that do not carry the virus so, over time, we can essentially make social contact free again, as it should be.<\/p>\n

    This crisis is unprecedented, and no single person or organization has all the answers. New perspectives often appear by recognizing and connecting patterns across silos. We\u2019ve been working with epidemiologists whose work focuses on the potential loss of lives. Economic models and scenarios highlight the loss of livelihoods. As we started discussing this with NGOs and global policymakers, we became aware of the synergies between these workstreams, and we sought an integrated perspective.<\/p>\n

    In the paper, we detail this framework and share data we have been collaborating on with policymakers. It is by no means perfect. We are already collaborating with a number of organizations on improving this work and getting additional data. Our hope is that, by publishing this work, we can invite others to contribute and leverage it so that we can bring more perspectives to bear on one of the great challenges of our lifetime.<\/p>\n