AI and COVID-19: How Artificial Intelligence is Being Used to Combat the Coronavirus
As Dave Chappelle once famously quipped, “Modern problems require modern solutions.”[1] Though this statement is rooted in comedy, it perfectly characterizes the 21st-century tool being used to address the COVID-19 pandemic: artificial intelligence.
South Korea was one of the first nations to channel artificial intelligence to combat coronavirus before there was even a confirmed case.[2] Seegene, a molecular biotech company, hosts an artificial intelligence-based big data system that helped the company quickly develop a test for coronavirus.[3] Seegene previously used its artificial intelligence system to make diagnostic kits for urethritis, but this was the first time the technology was used for a global pandemic.[4] These coronavirus tests require a chemical solution to be mixed with patient samples, which react in the presence of certain proteins.[5] Seegene owner Chun Jong-yoon attributes the faster and broader testing to its automatic nature, in which a diagnostic machine with a robotic arm mixes the solution and samples, performing numerous tests at once.[6] Notably, the United States has not used automatic testing for coronavirus.[7] Another benefit of using artificial intelligence to develop a test was that the company did not require a virus sample and only needed a few workers in its initial production; after government approval and South Korea’s high crisis alert, all Seegene workers were diverted to making test kits.[8]
Since Seegene’s development of this groundbreaking test, companies are using artificial intelligence to combat coronavirus in other ways. South Korean-based Deargen used a learning-based model and simplified chemical sequences to predict how strongly a molecule will bind to a target protein, and predicted an HIV medication to be most likely to block coronavirus proteins, as well as three other antivirals which might also block proteins.[9] Insilico Medicine, in Hong Kong, used artificial intelligence and a deep learning filter to generate and narrow a list of molecules that may bind to coronavirus proteins to prevent replication.[10] SRI International and Iktos, respectively based in California and Paris, have collaborated in efforts to develop anti-viral therapy by using a deep learning model to design virtual molecules and an automated synthetic chemistry platform to make them.[11] Finally, U.K.-based Benevolent AI’s learning model used data to isolate six compounds that block cellular pathways that allow the virus to replicate.[12] The use of artificial intelligence to combat coronavirus seems to be proliferating nearly as fast as the virus itself.
Another use of artificial intelligence includes predicting epidemics by analyzing sources and tracking infectious diseases.[13] For example, artificial intelligence company Blue Dot warned customers to avoid Wuhan before the CDC and WHO did.[14] Artificial intelligence is also used to lessen the burden on healthcare workers and remind citizens on care procedures.[15] In China, for instance, robots provide diagnostic checks, ambulances use artificial intelligence to weave through traffic, an app helps people determine if they were in contact with an infected person, and drones are used to ensure residents take proper precautions.[16] For example, predictions may become less specific as the disease spreads, and inconsistent data fed into artificial intelligence further compromises its predictions.[17] Additionally, artificial intelligence may not be ideal for early diagnosis, as physical signs of coronavirus may not show up immediately.[18] Finally, there is insufficient data on virus mutation, which inhibits artificial intelligence’s ability to predict evolution, and there are many potential drug candidates for treatment, which may take months to become viable.[19]
Artificial intelligence appears to be a promising, albeit imperfect, tool in combatting coronavirus. Despite its limitations, the coronavirus pandemic has ushered in a new era of medicine in which artificial intelligence can be used to mitigate effects and quickly develop treatment for viruses, and its usage will only grow in the future.
Footnotes
