- blog -

COVID-19 in the age of AI

Nobody expected the COVID-19 to have such a widespread and to cause such devastating effects. When the Spanish flu broke out in January 1918, the medics were not prepared for such a scale of illness. Two years later – in December 1920 – the number of infected oscillated around 500 million. In those years, unfortunately, medicine and technology were not developed enough to handle the pandemic outbreak, and artificial intelligence was relegated to a very few visionary science fiction feuilletons. Today, in the era of AI, we can not only gradually reduce the spread of the virus, but also recognize its symptoms quicker, and fight it effectively over time.

One of the first companies to register a very high risk of a viral infection spreading was BlueDot Global. This Canadian company warned, several months before the outbreak, the possibility of a pandemic using machine learning based on big data. Unfortunately, the human factor failed and no one took this warning seriously enough to introduce preventive measures. The virusbegan to spread to other continents using travellers as unknowing vessels, infecting more and more people.

Capturing the virus spreading is one of the main problems. Were it not for the solutions applied thanks to artificial intelligence, its expansion would probably look completely different. Many international, mainly SME, companies, using the computing power of their computers, began to implement various algorithms to track the expansion of the virus. For example, Follow @ home, changed its gaming computers into workstations to perform virus related calculations. In each millisecond, such computers are able to perform hundreds of billions of calculations and predict the next “virus steps”. Unfortunately, the calculation cannot prevent it from spreading, but can help to protect the most vulnerable sectors and areas and to contain it effectively.

COVID-19 detection tests are another huge problematic area that requires attention as it is critical to contain the spread of the virus and to provide immediate effective care. In Poland, the Institute of Bio-organic Chemistry of the Polish Academy of Sciences in Poznań, has prepared tests that are already in production and will allow to increase the test capability in the country. In Spain, the government decided to buy robots to perform automated infection tests. These robots can analyse each 80,000 tests in 24 hours. In Italy, the University Hospital Campus Bio-Medico developed a software with which, in 20 seconds, medics are able to find out if CT scans contain changes caused by COVID-19. Similar solutions are used and implemented by hospitals and biotechnology companies all over the world in the joint fight against the pandemic. The Polish company Intermedica has proposed a symptom checker, that, using machine learning strategies, increases the accuracy of symptom assessment, finding relationships and disease patterns in potentially infected people.

What can we do to reduce the risk of infection? Social distancing and special hygiene practices are one of the elements. However, a broader perspective and a long term strategy should be developed. Early detection of people with fever is one of the components that can give the first signal.

High hopes are also placed in the AlphaFold system, which, using deep learning algorithms, works on predicting the structure of the protein, contributing to a better understanding of the virus and to fight it more effectively. Currently, most countries join forces in the fight and strive to prepare effective drugs and vaccines. To reduce their development time AI techniques are effectively used to simulate the new drugs and to test them. Time is an extremely important factor because all substances must be refined quickly to be sure that they do not harm the patients while providing a quick response in the crisis.

In the meantime, hospitals and medical facilities, in addition to the standard protective measures, whenever possible, start using disinfection robots that can be operated remotely reducing the additional virus contact for the staff. An example of such robots are those from the company Blue Ocean Robotics or MTR, that are able to eliminate 99.99% of bacteria and viruses of the entire room in a quarter of the time human operators would need using, among others, UV lighting, disinfectants and special sensors for the detection of the infection level. Transport robots are another type of machines that can supports human operators. They limit the direct contact between potential infected subjects while ensuring the delivery of essential goods such as meals and medications.

How else can we monitor the virus spreading? A simple and effective technological solution is to install applications on phones, monitoring the number of people staying in quarantine, their location and the interactions between subjects to back trace eventual infection clusters. This would show virus concentration in the concerned areas and allow the virus tracing, monitoring and containment. This approach has some problematic issues as the GPS location and individual monitoring can be seen as a privacy invasion. AI would help to collect and process only data about the virus, not a specific person, still the matter is open and under discussion. It is important that the issue is solved fairly quickly to fight this pandemic are we up to the challenge?  There is no definite answer yet, because it is the first time in the 21st century that humanity has to deal with such a large problem.

The current situation is a global challenge that we are all trying to overcome together. Every day new data and new possibilities appear. Time is not our ally, but we believe that medicine supported by modern technologies and artificial intelligence will be able to cope with the current situation and to save human lives.