Written by: Naphat Permpredanun '24
Edited by: Megan List '24
Around a year and a half ago in January 2020, an emergency news report that a new virus had emerged in Wuhan, China changed everyone’s daily lives. The virus rapidly spread through airborne particles and had fatal effects. As the readers can probably guess, this virus is Coronavirus, or COVID-19 as we are familiar with it. This virus spread globally, infecting 200 million people and killing around 4 million people.
From these facts, if we turn back to just around a year ago, it’s reasonable that people thought that this virus might destroy humanity entirely.
But we know that this horrible fate has not occurred thanks to many recommendations from physicians and biology researchers, such as using masks, social distancing, and especially vaccination, that stop us from getting infected. As humans, we need to give a lot of thanks to the fields of medicine and biology for their help with the pandemic. However, if we say that only these fields helped us survive this virus, we risk underestimating contributions from other fields, one of them being Computer Science. One contribution from computer science is Artificial Intelligence (AI) , or the programming that is able to perform tasks that normally require human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages. This article will display some of the involvement that artificial intelligence had on the control of COVID-19.
The first contribution of AI to efforts against the coronavirus pandemic is called a “forecasting machine”, and it was used even before COVID entered the USA. It was first used by Boston Children's Hospital on the website HealthMap. They “[use] artificial intelligence (AI) to scan social media, news reports, internet search queries, and other data for signs of disease outbreaks''. Findings using AI were even faster than human’s discovery of this disease as shown by the finding that “Colleagues in Taiwan had already alerted Marjorie Pollack, a medical epidemiologist in New York City '' around a week after the machine was put into use but before the virus was in New York. Machine learning didn't stop there, since both HealthMap and other healthcare services tried to compete in predicting the outbreak of the Coronavirus.
The field of AI also alleviated the difficulties for curing and detecting the disease within the patients itself. This support could be divided into two parts : short-term curing (such as physical diagnosis) and long-term curing (such as accelerating the vaccination process, both scientifically and politically).
For short-term treatment, AI is used to diagnose whether the patient is infected by COVID-19 or not using the information of the CT-scan of lungs. Reading CT-scans takes a long time if done manually because there is a disparity between the result of the lung lesions of normal pneumonia and that of COVID-19. This made the diagnosis of COVID-19 a slow process and failed to catch the rapidly increasing amount of infectants. However, this problem could be solved by using the subtopic of AI called deep learning. Deep learning is the process of using computers that mimic human’s learning process, such that the computer will be trained by the pre-made training dataset, and then be used to help read CT-scans for COVID-19. By using this process, the time used to execute the accurate result for the CT-scan diagnosis decreases rapidly (from 2-3 minutes per diagnosis into a mere second), which helps to diagnose the rapidly increasing numbers of COVID patients.
In the long-term, AI also involves a process to create a vaccine. According to the National Center of Bioinformatics Technology (NCBI), the mRNA vaccines, like Pfizer or Moderna, are produced based on “the insertion of the encoded antigen in a DNA template from where the mRNA is transcribed in vitro” (NCBI). Therefore, the process must start from sequencing the proper mRNA in order to put it into a vaccine. However, this process takes lots of knowledge and time since only certain parts from the extracted coronavirus gene can be inserted into the vaccine. Therefore, scientists would have to use a time-consuming process to classify the sequence related to the vaccine by hand. This is when AI comes in. Within the capabilities of AI, there is a process called “supervised learning” which is a process that gives the computer a set of data (inputs) and the labels for the inputs (target) in order for the computer to learn. Then, in the future, this AI could classify more objects within its learning fields, leading to a faster classification process for the huge datasets. In this
Even though these technologies seem to one-sidedly be beneficial for researchers, some of the results from Artificial Intelligence could be harmful for both researchers and COVID patients. For example, according to Science.org, “From 2009 to 2015, Google ran an effort called Google Flu Trends that mined search query data to track the U.S. prevalence of flu. ‘At first the system did well, correctly predicting CDC tallies roughly two weeks ahead of time. But from 2011 to 2013, it overestimated flu prevalence, largely because researchers didn't retrain the system as people's search behavior evolved’. Searches for news reports about the flu were misinterpreted as signs of infection”. Considering this situation of flu prevalence and COVID-19’s to be similar in terms of people tracking news about the virus (which, in fact, may be in a higher degree than that of flu), this type of Machine Learning could make a huge blunder in the future.
CT-scan diagnosis and Gene sequencing could also cause a problem in terms of its accuracy , sometimes the number of examples are limited due to the number of cases available for study, which could create an inaccuracy. These inaccuracies affect every party, for example the programmers, physicians, and patients.
The other discussion based on the usage of Artificial Intelligence surrounds its ethicality. From the blog “Using AI ethically to tackle covid-19”, they classify the ethical concern into two main problems. First, the potential benefits increase the incentive to deploy AI systems rapidly and at scale, but also increase the importance of an ethical approach. The speed of development limits the time available to test and assess a new technology, while the scale of deployment increases any negative consequences. Without forethought, this can lead to problems, such as a one-size-fits-all approach that harms already disadvantaged groups. Second, the public may not fully trust in AI, which may prevent widespread acceptance of its findings. According to the BMJ, many people still don’t accept the usage of AI. For example, “concerns have been raised in China over the Health QR code system’s distribution of data and control to private companies”.
Within the catastrophe of COVID-19, artificial intelligence emerges as one of the best facilitators for physicians and researchers. It helped to detect the emergence of the disease, aided in vaccine development and allowed for quicker diagnosis of the disease. However, some concerns about AI are that it may inaccurately predict or diagnose COVID, and it may also be an overdeveloped technology that perpetuates bias of some groups. It may also take some time for the public to trust this technology. Nonetheless, this connection between AI and COVID-19 should be monitored and perhaps developed to tackle other diseases, too.
Cave S, Whittlestone J, Nyrup R, hEigeartaigh SO, Calvo RA. Using AI ethically to tackle covid-19 [Internet]. The BMJ. British Medical Journal Publishing Group; 2021 [cited 2021Oct24]. Available from: https://www.bmj.com/content/372/bmj.n364
CDC. Symptoms of COVID-19 [Internet]. Centers for Disease Control and Prevention. Centers for Disease Control and Prevention; [cited 2021Oct24]. Available from: https://www.cdc.gov/coronavirus/2019-ncov/symptoms-testing/symptoms.html
Cho A. AI systems aim to sniff out coronavirus outbreaks — COVID ... [Internet]. [cited 2021Oct24]. Available from: https://covid19.elsevierpure.com/en/publications/ai-systems-aim-to-sniff-out-coronavirus-outbreaks
Committee on Artificial Intelligence. Ai and control of covid-19 coronavirus [Internet]. Artificial Intelligence. [cited 2021Oct24]. Available from: https://www.coe.int/en/web/artificial-intelligence/ai-and-control-of-covid-19-coronavirus
Hosny A, Aerts HJWL, Ahmed Hosny. Artificial Intelligence in Medicine (AIM) Program Band WH, Hugo J. W. L. AertsArtificial Intelligence in Medicine (AIM) Program Band WH. Artificial Intelligence for Global Health [Internet]. Science. 2019 [cited 2021Oct24]. Available from: https://www.science.org/doi/full/10.1126/science.aay5189
Mariano B, Muneene D, Wu S. Artificial Intelligence for covid-19 response [Internet]. The BMJ. 2021 [cited 2021Oct24]. Available from: https://blogs.bmj.com/bmj/2021/07/23/artificial-intelligence-for-covid-19-response/
Zhang K, Liu X. Clinically Applicable AI System for Accurate Diagnosis, Quantitative Measurements, and Prognosis of COVID-19 Pneumonia Using Computed Tomography [Internet]. 2020 [cited 2021Oct24]. Available from: https://www.cell.com/cell/pdf/S0092-8674(20)30551-1.pdf
Randhawa GS, Soltysiak MPM, Roz HE, Souza CPEde, Hill KA, Kari L. Machine learning using intrinsic genomic signatures for rapid classification of novel pathogens: Covid-19 case study [Internet]. PLOS ONE. Public Library of Science; [cited 2021Nov28]. Available from: https://journals.plos.org/plosone/article?id=10.1371%2Fjournal.pone.0232391
Rosa SS, Prazeres DMF, Azevedo AM, Marques MPC. MRNA vaccines manufacturing: Challenges and bottlenecks [Internet]. Vaccine. Elsevier Ltd.; 2021 [cited 2021Nov28].
[Image Citation] Council of Europe. File: AI-COVID_870. JPG. Retrieved 2021, from https://www.coe.int/en/web/artificial-intelligence/ai-and-control-of-covid-19-coronavirus.