Silver Bullets from Silicon Valley: Artificial Intelligence will improve our ability to respond to global outbreaks like Covid-19

Is Covid-19 like MERS or is it like the pandemic Spanish flu? Which countries should impose travel restrictions? Are we over-reacting? Are we under-reacting? Where should the first batch of vaccines or antivirals be sent to? What is the best policy of containment? How will this affect global trade? Tourism?

We don’t know – and that is why we tend to be reactive. The public reacts to media. Leaders react to the public. The market reacts to policies. The media reacts to leaders and markets and so on and so forth.

While we may not have all the answers - we should be investing in global Artificial Intelligence (AI) systems that allow for more certainty in our decisions. These systems should have predictive analytics, integrate data from multiple sources, and be capable of recommending intelligent actions.

Imagining the AI System

For illustrative purposes, let us for a moment imagine how this system would look like.

First it needs basic information. This could include:

•  Real-time feed of all cases including demographics of patients, symptoms, and exposure (epidemiological links) & travel history. 
• Coefficients/Measurements of virulence and contagiousness. These would be variable and change in real time based on observed patterns and new scientific evidence like transmission vectors (insects, droplets, water, etc)
• Coefficients/Measurements that reflect “outbreak risk” for cities and towns. These could be based on population density, transport connectivity, and healthcare access and resources. These too could be variable and change as containment measures are put in place.

We can now begin to imagine a system that tracks patients in real-time and simulates millions of possible scenarios. As data comes in, it continuously readjusts its own parameters and recommends the most likely pattern for the outbreak.

That said, in some ways that is still a bit basic. Let’s add more information.

• The system scours the web for all scientific publications and begins to integrate new information on mortality and disease presentation. It begins to track different virus sub-strains and mutations.
• It then adds real-time GIS & traffic data form platforms like Google Maps and integrates inbound and outbound flight information. The system now begins to model outbreak risk and spread at a whole new level and starts to issue travel restriction recommendations.
• The system also links into key healthcare information like hospital occupancy levels, number of ventilators available, and stock levels of antivirals and gauges them against the model. It subsequently recommends resources to be sent well before they run out.
• It even integrates data from health apps that log self-reported fever and symptoms and is able to monitor the disease spread much faster than any single agency and offers ever more accurate predictive models.
• The system also is simultaneously analysing articles, news and social media posts and getting a feel for public sentiment. It recommends higher vigilance in complacent areas and highlights imminent public panic in others. It recommends that policies against panic purchasing and hording are put in place days before there is a run on grocery stores and shelves are emptied.
• The AI grows more and more sophisticated. It analyses auxiliary information like freight movement, port activity, hotel occupancy, public event calendars. Measures like split shifts, mandatory quarantines, and screening at airports are all integrated into the AI system as they are implemented. The system begins to offer indications on economic impact of the viral contagion itself as well as the measures put in place to contain it.

The Benefit

Lives are saved as resources and measures are used more effectively. Manufacturing and supply chains suffer less disruptions saving billions. The public feels calmer as officials back up their actions with real time data. People can even get personalized travel risk assessments before abandoning their holiday plans. Conferences and events are not needlessly cancelled. The market becomes more stable and feels more confident in the AI system’s predictions and recommendations.

What is Needed

Of course a system like this would need large scale information sharing at a global level. It would also need the buy-in of the tech giants like Google, Apple, and Microsoft in addition to the big social media platforms. It would also need a unique level of cooperation among healthcare providers, drug manufacturers, governments, transportation agencies, and many other stakeholders. That said, it is not as difficult as it seems. A great deal of the data already exists in the public domain and in terms of technology, many proto-systems are undoubtedly in development. Covid-19 offers us a genuine learning opportunity and I am confident that with the right mindset and investment, the next time we have a global pandemic threat we will have the right tools to guide decisions and ensure a more preemptive and effective response.

Feras Mahdi is an internal medicine physician, med-tech business leader, and AI enthusiast.