AI and the COVID-19 Pandemic
Updated: Dec 21, 2020
Emma Prévot analyses how and why Artificial Intelligence (AI) has played a vital role in the fight against the COVID-19 pandemic so far.
Artificial Intelligence was able to anticipate the outbreak of the novel coronavirus, predicting both the exact location where it originated and the first cities which saw COVID-19 cases.
AI spotted COVID-19 before everyone else knew about it
30th December 2019, a little after midnight, BlueDot, a Canadian AI start-up, spotted something unusual: 27 pneumonia cases associated with a market selling seafood and live animals in Wuhan, China. BlueDot soon flagged what would come to be known as COVID-19, nine days before the World Health Organization (WHO) released its first alert statement.
Let’s dive deeper into how this artificial intelligence platform has been able to anticipate the outbreak of the pandemic.
BlueDot works with big data; as a matter of fact, it’s able to process an incredibly large amount of data (more that 100,000 online articles each day, in 65 different languages) using natural language processing and machine learning. Data comes from sources like global airline ticketing data; public health organizations; human, animal and insect population data.
The AI findings are then reviewed by physicians who decide whether to send an alert to authorities or not.
In the case of COVID-19, not only BlueDot was able to send an alert of the novel coronavirus, it also correctly anticipated where the infected might be travelling, predicting 11 of the cities which first saw COVID-19 cases.
Artificial Intelligence has been vital to arm ourselves against the virus. Several are the areas in which it has contributed, ranging from contact tracing and prevention maps to early detection and quick diagnosis.
AI has contributed in the fight against COVID-19
Not only was AI able to anticipate the outbreak of the novel coronavirus, but it also promptly started to help our fight against it. From the very beginning of this pandemic, Artificial Intelligence started working behind the scenes to try to fill the lack of human expertise and knowledge and to develop tools to arm ourselves against the novel coronavirus.
In April 2020, EIT Health started a research named ‘Digital Control Centre for COVID-19’. The aim of this project was to demonstrate that the use of artificial intelligence tools can lead to a significant reduction in mortality rates amongst hospitalised COVID-19 patients. The AI solution consists of a virtual control centre for infected patients, under the supervision of an expert infectious disease specialist. Combining data collection and clinical guidance, it allows to independently assess and/or change a patient’s treatment, identify and prioritize the most seriously ill ones, tailor specific therapeutic approaches and decide when to discharge a patient.
The research was conducted on 786 patients admitted to Hospital Clinic, Barcelona, and led to a 50% reduction in mortality rates.
The solution is now being validated and expanded to other hospitals in Spain, the Netherlands and Belgium.
Artificial Intelligence and Machine Learning have been at the core of the strategies adopted across the globe to slow down the spread of the virus. There are three macro areas where these modern technologies have concretely come in action:
1. Early detection of the infection
Frontline physicians and the healthcare system have been, from the very beginning, under considerable stress to test and treat patients. HealthTech companies came to the aid to develop AI based algorithms that could quickly analyse enormous quantity of medical images (like Computed tomography (CT) or Magnetic resonance imaging (MRI)) to locate the virus in patients’ lungs, and distinguish it from other respiratory infections.
This cost-effective process has enabled faster decision making, reducing the burden on medical practitioners.
The Chinese tech giant Alibaba, for instance, developed its own AI system which can detect the virus in only 20 sec with 96% accuracy. For comparison, it takes 15 minutes for doctors to make a determination).
2. Containing the spread
Artificial Intelligence excels in analysing and uncovering patterns in data, which makes it a perfect tool for tracking the spread of the virus. Many different solutions and approaches have been adopted all over the globe, but two of them proved to be the most effective:
3. Contact tracing apps
According to current estimation, the average number of cases a single case generates (i.e. the number of people infected by a person positive to COVID-19), is between 1.4 to 2.5. Being able to easily and quickly identify these persons that have been exposed to the virus, is key to slow down the rate at which the pandemic spreads. This process is called contact tracing.
Various infected countries tried to digitalise this contact tracing process using mobile applications that citizens need to download. These apps are designed to collect individual personal data, which are then analyzed by ML and AI tools to trace a person and the people they have been in close contact with. Eventually, if a user tests positive, the app will notify others who have recently been nearby, and tell them what to do.
The effectiveness of these apps relies on the number of people that downloaded them; a study from the University of Oxford showed that to bring an outbreak under control, at least 60% of the population have to enter a contact tracing system. Not even one country has taken-up this threshold yet.
Nevertheless, the situation improved after a surprising collaboration between two Tech Giants: Apple and Google.
Their joint effort led to a Bluetooth-based contact tracing system, directly built into their platforms (iOS and Android), with no external app requirement. Phones would anonymously keep track of other phones that were nearby. Users are still required to opt in for privacy reasons but they don’t have to download any external app.
- Prevention maps for identifying coronavirus hotspots
Facebook has been one of the first to exploit the power of AI: using anonymized data about people’s movements, it created accurate population density estimates, to understand when, where and with whom, people were moving and congregating.
This information has been essential for researchers to forecast how COVID-19 was spreading, to understand where stricter policies to reduce the proliferation were needed, and to allocate medical resources more effectively.
In general, AI can identify the most vulnerable regions, people and countries so that authorities can take measures accordingly.
- Development of drugs and vaccines
AI algorithms keep analysing the available data on COVID-19 to help in designing new drugs, and finding whether drugs previously used for other treatments, could also be an effective cure. One of the main reasons why AI research has proven to be more efficient than human research is because it can speed up drug testing processes and clinical trials during the development of a vaccine. Furthermore, AI is being used to suggest components of a vaccine by understanding the viral protein structures.
Initially, AI was designed to overcome simpler problems, like winning a chess game; as of now, with the technological advancements, AI has become more and more sophisticated, solving complex problems more efficiently, rapidly, and at a lower cost.
Even though a pandemic wasn’t exactly what AI had been trained for, its role is now vital and is rightly used as a tool complementing human intelligence.
If on one hand the COVID-19 outbreak is still putting the global healthcare system under immense stress, on the other it is boosting a drastic change in Medicine. Innovation can lead to a brighter future in the field, concerning AI based faster and more accurate diagnosis, broader medical support, drug discovery and more efficient prevention. In the near future, companies such as the aforementioned BlueDot will prove more and more essential in warding off future coronavirus pandemics and venture capital is going to pour into healthcare to provide fresh impetus and motivation for brilliant and smart minds.
The UCL Finance and Technology Review (UCL FTR) is the official publication of the UCL FinTech Society. We aim to publish opinions from the student body and industry experts with accuracy and journalistic integrity. While every care is taken to ensure that the information posted on this publication is correct, UCL FTR can accept no liability for any consequential loss or damage arising as a result of using the information printed. Opinions expressed in individual articles do not necessarily represent the views of the editorial team, society, Students’ Union UCL or University College London. This applies to all content posted on the UCL FTR website and related social media pages.