Can Artificial Intelligence Make a Difference In Combatting Coronavirus?

The mysterious coronavirus has been spreading quickly in the last few weeks. There has been a minimum of 305 fatalities and 14,300 people have the infection. (

Last week, the World Health Organization (WHO) stated that the coronavirus is a global emergency. To give this perspective, it has already gone over the numbers of infects people during the SARS outbreak of 2002 and 2003 in China.

Many countries are working hard to put a stop to the virus. There have been several quarantines, lockdowns on large cities, travel limits, and faster research on vaccine development.

But is it possible that artificial intelligence (AI) could lend an assist? It actually already has.

BlueDot Is an AI Firm Already Helping With Pinpointing Virus

One example is BlueDot, which is a new venture-backed startup. The firm has designed a complex AI platform that processes billions of data pieces, such as from the air travel network around the world, to look for outbreaks.

With the coronavirus, BlueDot gave its first alert on December 31. This was before the CDC in America, which made an alert on Jan. 6.

BlueDot was the idea of Kamran Khan, an infectious disease doctor, and professor at the University of Toronto specializing in Medicine and Public Health. He was a mainline healthcare worker during the SARS outbreak 18 years ago.

Khan noted recently that his company is using natural language processing or NLP and machine learning to process large amounts of text data that is unstructured. It is currently processing 65 languages to track possible outbreaks of 100 disease, every 15 minutes, 24 hours per day. If this work was done by hand, it would probably require 100 people to do it right. The advanced data analytics allows health experts to focus their precious time and energy on how to respond to disease risks, instead of spending time on gathering and organizing vast amounts of information.

BlueDot will probably not be the only company to leverage AI to curb coronavirus. Here are other possibilities on the horizon:

Colleen Green, GM of Healthcare, DataRobot

Greene believes AI could predict many new cases by geographic location and which types of populations could be at the highest risk. This sort of technology may be used to warn travelers so that the most vulnerable populations can wear medical masks while they are on the road.

Vahid Behzadan, Computer Science Professor at the University of New Haven

AI can help with the improvement of optimization strategies. For example, some recent research on the use of machine learning to optimize social distancing or quarantining between cities, communities and countries to control the spread of epidemics could be groundbreaking. Also, my clinical research group is working with others in the field to develop methods to enhance vaccination strategies that leverage recent advances in AI, especially in reinforcement learning techniques.

Dr. Vincent Grasso, IPsoft Global Practice Lead for Healthcare and Life Sciences

When a disease outbreak happens, it is important to gather clinical-related information from patients and others that are involved, such as physiological state before and after the outbreak, logistical information about sites of exposure, and other vital information pertaining to the outbreak.

Deploying humans into such situations is difficult and expensive, especially if there are several outbreaks, or if they are in countries that do not have enough resources. Conversational computing can work as an extension of humans to get important information and would be a fantastic addition. Conventional computing can be bidirectional in that it can work with a patient and collect information. Or it can do the reverse – provide information based on plants that are modified or standardized based on assorted situational variations.

Also, working in a multilingual and multimodal manner would extend the conversational computing deliverable. In addition to this type of front end benefit, the data being collected from several sources such as text, voice, and medical devices and others are highly beneficial data points that can help doctors to learn how to fight an outbreak in the future more effectively.

Steve Bennet, Direction of Global Government Practice at SAS

AI can help with the coronavirus in several ways. It can predict hotspots around the globe where the virus may jump from animals to humans, known as a zoonotic virus. This can happen at exotic food markets that have poor health codes. Once an outbreak has been found, health officials can engage with artificial intelligence to learn how the virus is going to spread based on the conditions in the environment, people’s access to healthcare, and how it is being transmitted between people.

AI also can pinpoint and find commonalities within the localized outbreaks of the virus. Or, it can use micro-scale adverse health events that are unusual. The insights from such events can answer many of the unknowns about the nature of this virus.

When it comes to finding a cure for this virus, creating vaccines and antivirals is a long trial and error process. But the medical community has been successful in finding several vaccines for similar viruses in past years. So, using AI to find patterns from similar viruses and detecting attributes to look for to build a new vaccine gives medical researchers a better chance of getting lucky than if they were building one from scratch.

Don Woodlock, VP of HealthShare at InterSystems

As machine language improves, we can read billions of data points and clinical documents in patients’ medical records and find connections to patients that do or do not have the disease. The features of each patient that get the disease will pop out of the process of modeling, which can help us to find patients that have a higher risk.

Also, ML techniques can build a model or relationship between treatments that have been documented in medical records and later patient outcomes. These models can identify treatment options that are related to better outcomes and help guide the process of making effective clinical outlines.