By now, we all know COVID-19 is spreading fast, infecting people at an exponential rate. “This is the defining global health crisis of our time,” says Dr. Tedros Ghebreyesus, Director-General of the World Health Organization. “The days, weeks, and months ahead will be a test of our resolve, a test of our science, and a test of solidarity.” 

As the world braces for new drug treatments and vaccines, researchers across the globe are turning to emerging technologies like Artificial Intelligence (AI)  to identify and track the spread of COVID-19. The same AI technology used in big data and cybersecurity measures is now being deployed as a tool for public health. 

AI had a central role in first identifying the COVID-19 outbreak, and forecasting its spread across the globe. Currently, it’s now being used to help diagnose and potentially develop treatments for the outbreak as the pandemic accelerates across the globe. Let’s begin by learning how AI predicted the COVID-19 pandemic in the first place. 

Assessing the outbreak

On January 9, 2020, the World Health Organization alerted the public about a flu-like outbreak in Wuhan, China. A cluster of pneumonia cases were reported, possibly from human exposure to live animals at the Huanan Seafood Market. Ten days earlier, BlueDot – an AI startup based in Canada – sent alerts to its clients in the public and private sectors about a possible outbreak on December 31, 2019.

According to the company’s website, BlueDot “uses big data analytics to track and anticipate the spread of the world’s most dangerous infectious diseases.” Using natural-language processing and machine learning techniques, the AI-driven algorithm analyzes billions of data points including global news reports, airline data, and reports of animal disease outbreaks, to understand the spread of COVID-19. 

Once the data-sifting was completed by BlueDot’s AI, epidemiologists assessed the conclusions before reporting it to government, business, and public health clients. Kamran Khan, the founder of BlueDot told the Canadian Press, “We happen to have growing access to data we can use… to generate insights and spread them faster than the diseases spread themselves.”

With commercial air travel, humans are natural virus carriers, dispersing regional diseases like COVID-19 around the world. Using airline travel data, BlueDot maps how viruses spread through flights out of Wuhan and other global transportation networks to predict countries and regions most at risk for an outbreak. 

With infectious disease pandemics like COVID-19, speed matters especially when countries like China are hesitant to share information in the early days of an outbreak. By sifting through public data sets like news reports, discussions on online forums, and airline data –  AI has demonstrated how it can predict outbreaks faster than traditional channels. 

Tracking how the virus spreads

Travellers at Bangkok International Airport walking with luggage, wearing surgical masks following the Coronavirus outbreak.

By February 2020, WHO officials were relying on an AI system called HealthMap from Boston Children’s Hospital, to track the global spread of the disease. Using publicly available data on the Internet HealthMap “monitors, organizes, integrates filters, visualizes…early detection of global public health threats.” This information is shared with health agencies like the World Health Organisation to help them decide where they should focus their immediate efforts. 

“What’s really phenomenal here is we’re seeing incredible international collaboration and a huge amount of data sharing” says John Brownstein, professor at Harvard Medical School on ABC’s news podcast The Signal. The HealthMap website is being updated in real-time providing visitors the ability to animate the global spread of coronavirus from its early detection in Wuhan, China.  

COVID-19 diagnosis

coronavirus screening at medical centre

In the early days of the outbreak, patients were overwhelming Chinese hospitals with pneumonia-like symptoms. Software from Beijing startup Infervision, was deployed in 34 hospitals in China to expedite diagnosis on more than 32,000 potential cases. 

Infervision’s AI was originally designed to diagnose lung cancer from CT images. By mid-January, Infervision realized that existing customers were employing a little used feature in the software that looks for evidence of pneumonia. During the Lunar New Year holiday, Infervision’s staff tuned their algorithm with more than 2,000 images from COVID-19 patients. 

With this new data set, hospitals were using Infervision’s AI to flag patients that may have the disease from CT images. Doctors would follow up with other examinations and lab tests to confirm diagnosis, eliminating a significant bottleneck to quickly identify most at risk patients. 


Group of medical researchers analyzing medical data on a computer in the laboratory.

Outside of a diagnosis tool, AI is being employed by research teams to accelerate discovery of drug treatments and vaccines.  One method is running simulation tests to screen through millions of chemical compounds for potential drug discoveries. AI can perform this task far faster than human experts, which helps researchers narrow down chemical compound options for testing and trials. 

The second method is identifying targets that new drugs can latch onto to slow the spread of COVID-19 between people, and reduce the impact on those infected.  Google’s DeepMind – known for being the first computer program to defeat a professional human Go player – is focusing on the second method. The DeepMind team is hoping AlphaFold, its deep learning system, will contribute to the understanding of how the virus functions by “predicting protein structure accurately when no structures of similar protein are available.” 

This “free modelling” approach uses deep learning systems and newly developed methods to provide researchers with predictive structures of some of the understudied proteins in SARS-CoV-2. With the help of AI and new machine learning methods, researchers can quickly identify certain proteins that could become potential drug or vaccine targets. 

AI’s role in managing COVID-19

Without question, artificial intelligence has played a pivotal role in detecting the COVID-19 outbreak, tracking its spread across the globe, and streamlining early detection. As the number of cases and death toll rises, so too does the supply of data that can be shared across global AI research teams. The hope is that emerging technologies like AI will help researchers find breakthrough drug treatments and vaccines for COVID-19. 

Nobody knows for certain how this pandemic will play out. Nevertheless, we should all take comfort knowing that the brightest medical researchers, computer scientists, and AI are all collaborating to tackle what Bill Gates describes as a “once-in-a-century pathogen.”

The post How AI might help us diagnose and track COVID-19 appeared first on Futurithmic.