Sorry, you need to enable JavaScript to visit this website.

Is COVID-19 the watershed moment for Artificial Intelligence?

Is COVID-19 the watershed moment for Artificial Intelligence?
May 05, 2020

Technology and digital advancements are best meant for world sustainability, be it for the alleviation of poverty, the reduction of inequalities, or bringing about gender equality. Technology, when adopted by governments around the world for addressing the 17 sustainable development goals that the UN charter had ratified in 2015, is the best form of digital transformation to have.

“Good Health and Well-being” is one of those 17 sustainable development goals too, and one can’t help but wonder if the advancements in technology and the rapid strides being made in digital transformation can in some way help alleviate the dangers that the world faces today.

Artificial Intelligence and COVID-19

Since December 2019, when the unrelenting coronavirus disease, or COVID-19, was identified in China, artificial intelligence (AI), machine learning, data and analytics, and natural language processing, have been clamored about as a few potent solutions we have to fight the disease and bring about sustainable development.

And rightly so. On December 31, 2019, the Canadian AI company BlueDot rose to fame when it alarmed its customers about the pneumonia cases that had surfaced a few days back in Wuhan, China. This was nine days before the World Health Organization confirmed the finding of the coronavirus.

Therefore, AI, with its combined might of machine learning, natural language processing, and computer vision had announced its potency right since the advent of the disease. AI models have the potential to contribute toward the combined fight through each step of the “containment chain” namely- TRACK, TREAT, ANALYZE, PREDICT, and DEVELOP.

The graphic below gives a snapshot of the AI possibilities in a pandemic world:

pandemic world


The most unsettling aspect of COVID-19 has been the manic speed at which it spreads. Therefore, most countries across the world have announced quarantine measures as a local outbreak would wreak unstoppable mayhem.

COVID-19 could be a watershed moment for artificial intelligence as well as the data and analytics ecosystem

CCTV Surveillance

That is where Neurons Lab, an AI consulting boutique, has analyzed CCTV camera footage and applied machine learning, data and analytics, deep learning, and mathematical modeling to arrive at upto 56% accurate detection and tracking of potential COVID-19 patients.

Biosignal Measurements

Likewise, AI is capable of detecting heartbeat fluctuations from a change in facial color or even by the involuntary upward movement of the neck and head. Nuralogix, an AI startup has released an AI app called Anura which is capable of estimating blood pressure fluctuations by recording a 30-second video selfie.

Wearables Data

The signals emitting from our pulse rate through fitness trackers can speak volumes on our physical and emotional state and researches have shown that prolonged deviation of vitals such as blood pressure can hint at an early onset of flu or may be very early hints of the typically asymptomatic coronavirus.

Lack of enough data points has disallowed AI to completely blossom here.


Quick and precise diagnosis of COVID-19 can not only stall the spread of the infection but also generate valuable data to train the AI models. Researchers at UN Global Pulse opine that AI, with its accurate computer vision, can perform diagnosis faster and better than radiologists (who, being human, might get fatigued and err on judgement).  This is good news at a time when COVID-19 test kits are in short supply and the application of computer vision on X-Rays and CT scans can provide for early diagnosis, treatment, and overall control on the disease.

The potential of AI in treatment has barely been used, although it has been reported that a number of Chinese hospitals have deployed ``AI-assisted" radiology technologies.


In an age of social media and digital transformation, there can’t be a dearth of data. While many experts question its quality and call it adulterated enough for it to not be considered for any structured analysis, nothing beats the medium if the whole purpose of a study is to analyze the broad patterns of the disease.

John Brownstein, the Chief Innovation Officer at the Harvard Medical School, is an expert on gleaning social media for health patterns and scours through social media posts, news reports, data from official public health channels like WHO, and information supplied by doctors for getting warning signs on the virus across the world.

His team searches for mentions of specific symptoms from a location where lots of reported cases of the infection has come out. Natural language processing is then deployed to read between the lines and separate a news post from a complaint by a potential virus victim on their symptoms.

The Canadian AI startup BlueDot which was mentioned at the beginning of this article, goes beyond social media data to track the worldwide movements of more than four billion travelers on commercial flights every year; population data, climate information; and local information from journalists/healthcare workers, rummaging through 100,000 online articles each day spanning 65 languages, making their work immense in scope.


Such analysis (though on unstructured data) allows them to undertake statistical analysis and draw out predictions for the future (a sample of one such statistical modeling is shown below)

pandemic world

Such predictions can also help in calibrated and calculated lockdowns, instead of blanket shutdowns that we see today. Areas where the pattern shows a drop can look at opening up people and economy in a calculated manner and only those areas where the predictions aren’t palatable might consider for a lockup extension.

Such mathematical models are still at a very nascent stage as the quality of data from social media (though huge but lacks quality) is driven more by bias and hubris and is far away from practicality.


Finally, an area where AI holds great promise but is at a very nascent stage is the ability to create vaccines at scale. Google’s DeepMind has predicted the structure of the proteins of the virus that could help in discovering effective vaccines and drugs for COVID-19.

Likewise, researchers at Benevolent AI, a UK-based AI startup, have identified Baricitinib, a drug used for rheumatoid arthritis and myelofibrosis, as a potential treatment for COVID-19.

Similarly, a Singaporean firm Gero used a deep neural network to identify a number of existing and approved drugs that could potentially be used to treat COVID-19.


Artificial intelligence holds enormous potential to sustainably contribute toward global healthcare at large and pandemics like COVID-19 in particular. Across the pentagonal opportunity spectrum of Track, Treat, Analyze, Predict, and Develop, AI has multiple use cases but the unavailability of quality data (biased data on social media and reluctance of world bodies to share public health data and records for privacy and compliance reasons) has led to slowness of experiments and results. This makes AI investment and AI research paramount.

But COVID-19 has changed all of that. Data scientists are getting more and more ad-hoc requests to create models around COVID-19. Public health bodies have become more open to share health data and patient records as long as the intention is noble. Governments are beginning to leverage AI firms for strategic decision-making.

Thus, we see a huge spurt in AI investments into healthcare and AI research in general. COVID-19 could be a watershed moment for artificial intelligence as well as the data and analytics ecosystem.

While this was an experiment, am sanguine that come the next pandemic, our data scientists and AI firms will be standing as the first layer of defense alongside the brave epidemiologists the world over.