The COVID-19 outbreak has rocked the globe, with many countries going into a complete lockdown. This article takes a look at how AI can be humanity’s best hope in containing the pandemic as well as preventing and containing such outbreaks in the future.
The COVID-19 virus has affected the entire world in a profound way. For the first time in living memory, most of the airports in the world are shut and many countries are in a state of a complete lockdown. While the death toll from the virus may not be as disastrous yet, as compared to previous epidemics and pandemics, the sheer number of infected people is staggering. At the time of writing this article, over 580,000 people have been infected by the disease. Over 26,000 people have lost their lives. In Italy alone, over 9000 people have died. Over 100,000 people have been infected in the USA making it the worst hit in terms of the number of COVID-19 cases. Scientists predict that it is going to take a few more weeks for the disease to peak in most countries. This could possibly mean that the numbers quoted above will only increase in the coming days, reaching new heights, before they start to fall.
Why AI was not very effective in preventing COVID-19 from turning into a pandemic
A huge number of AI projects, both open source and proprietary, have been focused on revolutionising the medical industry for quite a few years now — be it diagnosing a disease, building smarter testing equipment or creating smarter mechanisms for delivering medical aid to places the medics might find difficult to reach.
However, all this research doesn’t seem to have made much of an impact in preventing the spread of COVID-19. One might be inclined to question the efficacy of AI itself at this point. However, being an AI researcher who has closely followed the developments in the AI community during this crisis, I have reasons to disagree with that notion.
On January 6, 2020, the Centre for Disease Control (CDC) of the USA issued a travel advisory for people travelling to China regarding the COVID-19. This was followed by the WHO’s advisory on January 9. However, an AI powered health monitoring system based in Canada named BlueDot had issued an advisory regarding this outbreak on December 31, 2019, a full week before any government body or health organisation was able to predict the outbreak. What makes things more interesting is that the predictions that BlueDot made regarding the spread of the virus turned out to be very accurate. BlueDot uses natural language processing and machine learning to scour thousands of news reports across the globe, as well as airline ticketing information, to correlate and predict the spread of an epidemic/pandemic. This is a much superior approach as compared to trusting the official information shared by a country’s government, which may not always be accurate or provide a complete picture on the ground.
The example of BlueDot proves that it is not the efficacy of AI that is to blamed but the delay in our response to the pandemic. Simply put, we were slow on pulling the trigger on this one. Nobody expected COVID-19 to spread as fast as it did and have such a major impact on the entire world. So, we were slow to act. Not enough of an effort was made during the months of January and February. By the time people started to realise the potential of the threat that COVID-19 posed, March had arrived and the disease was in full swing, causing countrywide shutdowns and chaos.
How is AI currently helping with the COVID-19 pandemic?
While AI may not have been very useful in preventing the pandemic from occurring, it is certainly playing a role in containing the pandemic and helping medical teams in treating the infected patients.
- Facebook is currently working with researchers at Harvard University’s School of Public Health and the National Tsing Hua University, in Taiwan, sharing anonymised data about people’s movements and high-resolution population density maps, which help them forecast the spread of the virus.
- Google is using machine learning and AI to fish out the misinformation regarding COVID-19 on its platforms.
- Exscienta, a UK based pharmaceutical company, is using AI to develop a drug to fight against the spread of COVID-19.
- Deep Mind, a Google owned AI based company, is using its AlphaFold system to release structure predictions of several proteins associated with the virus.
- Companies such as BlueDot and Metabiota have been key players in predicting the growth of the pandemic, which is vital information for both the governments and the medical staff.
- Alibaba Cloud, DAMO Academy, and DingTalk have joined forces to launch a series of AI technologies and cloud based solutions to support companies and organisations worldwide in the fight against COVID-19.
- Researchers in the field of AI from the Renmin Hospital of the Wuhan University have developed an AI solution that detects COVID-19 in patients with 95 per cent accuracy, by analysing the CT scan of a patient’s chest.
- The Israel based medtech company, Nanox, has developed a mobile digital X-ray system that uses AI cloud based software to diagnose infections and help prevent epidemic outbreaks.
Open source projects in the fight against COVID-19
While the big tech companies and research institutes have been doing their part, many open source projects have also been started, using the power of AI in the fight against the COVID-19 pandemic.
CHIME, a COVID-19 hospital impact model, has been made open source by the Predictive Health Care team at Penn Medicine. CHIME uses SIR modelling to predict the number of people that can be infected in a closed population. This tool is proving to be very useful for hospitals to predict and plan the number of beds required to treat the patients.
NEXTSTRAIN is a real-time tracking program for pathogen evolution which is currently being used to monitor COVID-19. This technology, built by the Global Initiative on Sharing All Influenza Data (GISAID), had previously proven to be very useful in the fight against Ebola.
In India, CoronaSafe is an open source initiative launched with the support of the Kerala government to combat COVID-19 on multiple fronts. The Corona Literacy Mission, Corona Care Centres and real-time updates of COVID-19 cases in India have been made available on this platform. A wide array of scientists and engineers from multiple disciplines, including AI, are participating in this project.
The OpenCovid-19 initiative launched by Just One Giant Lab (JOGL) is working on building tools and techniques to fight the COVID-19 pandemic.
COVIDathon is a hackathon being conducted by the Decentralized AI Alliance and its partners SingularityNet and Ocean Protocol. The eight-week online hackathon aims at bringing AI enthusiasts and experts together on a single platform to build AI tools that will help in containing and managing the COVID-19 pandemic.
The role of AI and FOSS in preventing and containing pandemics in the future
As far as the prevention of a pandemic is concerned, a major lesson that the COVID-19 outbreak has taught us is to never underestimate the speed with which a contagious disease can spread. COVID-19 has disproved all predictions about the number of people it would infect, bringing entire countries to a halt and battering the global economy in the process. Speed is everything when it comes to preventing epidemics and pandemics. In the age of Big Data, only relying on statistics provided by government officials is no longer an option. Only sentient machines powered by AI have the required capability to scour through massive amounts of unstructured data across multiple mediums on the Internet to reach a prediction fast enough to prevent an epidemic. Therefore, it is beyond a doubt, that AI will play a major role in preventing the spread of diseases in the future.
There are two major challenges that we face in containing a contagious disease after it spreads. The first major challenge is social distancing. While social distancing can be voluntarily done to a certain extent, activities such as grocery shopping and making runs for essential supplies, force a person to come into contact with infected individuals. AI powered drone technologies can be a major asset in this regard. Automating the collection and delivery of essential supplies using drones can make social distancing truly effective for people, especially when we are faced with lockdowns and curfews. It will also help in preventing the unnecessary hoarding of goods done by certain individuals during times of crisis, thus reducing shortages of key supplies.
The second major challenge is that members of the medical community are forced to come into contact with infected individuals during a pandemic, thus risking their lives and the lives of their families in the process. Major strides are currently being made in the field of robotics to remedy this problem. However, in future, more effort and funding will be required for building sentient robots that are able to provide at least the basic medical care to the patients, such as checking their vital parameters, taking basic tests, performing diagnosis and cleaning the premises. This will reduce the risk to doctors and the hospital staff as well as help with containing the disease.
I would like to conclude this article by saying that epidemics and pandemics affect us all. It is not the problem of some but the problem of everyone. This requires us all to pitch in and make an effort in preventing and containing such outbreaks. It is a matter of great importance that the open source community makes an effort in furthering the research and development of the AI technologies that will prevent such a situation from occurring again in the future. The beauty of open source is that it not only gives everyone access to pitch in, but also provides everyone with the freedom to use the technology for free. In the under-developed and developing countries, such technologies can be crucial in saving the lives of millions of people.