Five Ways How AI Can Fight The Spread of COVID-19 Pandemic

The coronavirus pandemic is an example of a “black swan”, an event that no one could predict, that is a game-changer for multiple industries and affecting the whole world (although Bill Gates would probably disagree…).  

Almost all the global know-how and various technologies have been harnessed to tackle the crisis, including machine learning and artificial intelligence. These are powerful tools that WILL help to tackle the threat and minimize the impact.

The black swan theory mentioned above refers to an event that is nearly impossible to predict and has a tremendous impact on the whole world. The key aspect of the black swan is the fact that there are: 

• The disproportionate role of hard-to-predict and rare events that are beyond normal expectations in history, finance, science, and technology. So, although there has been an epidemic before, it is nearly impossible to spot one in advance. 
  
  
• The non-computability of the probability of consequential rare events due to the small probability of one.
  
  
• The psychological biases blinding people, both individually and collectively.

The coronavirus pandemic is a gold standard of the black swan. But it is not making the event easier to manage. 

The COVID-19 pandemic - a handful of facts

The ongoing pandemic has been first spotted by China and reported to the WHO country office on 31 December 2019. The outbreak was declared a Public Health Emergency of International Concern on 30 January 2020. The new disease attacks the respiratory system, being particularly dangerous for people of age and weakened by additional diseases. 

The key challenge in managing the outbreak is the time between the exposure and symptoms onset which is typically around five days but may last up to 14 days. Also, the disease produces symptoms easily mistaken with the common flu or cold, including fever, cough, and shortness of breath. 

Considering the lack of a cure, there is only symptomatic and supportive therapy available. The easiest (not to be mistaken with easy!) way to stay healthy is to avoid interaction with other people, avoid touching face and keep hands clean. 

According to the UN Secretary-General Antonio Guterres, the outbreak is apparently

“the most challenging crisis we have faced since the Second World War”

“represents a threat to everybody in the world and an economic impact that will bring a recession that probably has no parallel in the recent past”. 

Considering that, it is high time to get on the digital hands of artificial intelligence to work.

How AI can help to fight COVID-19 

To be honest, the fight with the disease is about the people on the front line - brave healthcare employees, including physicians, paramedics, and nurses. The AI technology can perform only an auxiliary function - but sometimes it can be enough.

Creating smart dashboards

In the age of big data, access to the latest and freshest information is the first step in making more informed decisions. Delivering a smart dashboard about the ongoing situation can be a blessing for the decisive people. 

The AI-powered dashboard can deliver nearly any information in a comprehensible and easy form, be that a geographical placement of sick and quarantined people, the distribution of the medical equipment, or the capacity of medical staff and infrastructure. 

The smart dashboard can also be augmented with the forecasting capabilities not only to deliver current information but also how it is predicted to change in the near future and suggest what to do - it is better to deliver masks before the medical staff needs to improvise or worse - to work with no protection at all. 

X-ray and medical images analytics

According to the IBM data, up to 90% of all medical data today has a form of an image. Also, contrary to the image of a cat, dog, or many other types, all these images are already provided with a standardized description - labeled, a data scientist would say. Thus, it is a matter of training a proper neural network on a prepared dataset to get a machine that supports the diagnosis by preprocessing medical images. 

When it comes to the coronavirus outbreak, it can be automated analysis of the X-ray of lungs, that can deliver a significant piece of information on the patient’s condition. One of the most common complications after the COVID is respiratory diseases including pneumonia. 

Also, the blood sample visual analysis can provide additional information on a patient’s health and overall condition. 

Handling the documentation

Medical data is Big Data - including the need to fill in the documentation and search the databases for knowledge on the patient and his or her condition. The need to fill in tons of documentation can stack up the medical workforce and not only. 

The need to deliver the documentation about the remote work of teachers and other specialists who were suddenly transferred into their homes and asked to work are also overwhelmed by the need to document their work and show progress. 

Considering that, the AI can support them by delivering the automated analysis and providing more detailed information about their performance. For example: 

• The physician can get all the information about the patient straight to his or her tablet device and start treatment with no need to dig up the paper archives. 
  
  
• A teacher can get his work tracked and analyzed in real-time, including the feedback collected from students. 
  
  
• Public service providers, transferred into remote handling of their duties, can benefit from getting their messages preprocessed and transferred right to the destination department, eliminating the bottlenecks and limitations of manual handling of the incoming information. 

These are only a few examples - there are myriad of ways how AI can support business operations, but these three have been chosen due to the sudden and unexpected workload transformation in the organizations mentioned above. 

Image recognition - masks, isolation

Yet another way to use image recognition technology is to harness the power of AI to encourage people to wear masks. According to The Centre for Disease Control and Prevention every American should wear a mask to defeat a Coronavirus.

Facial recognition technology is currently well developed and the technology has little to no problem in spotting the difference between shapes of eyes, noses, and mouths. Thus, it would be not a challenge to deliver a model that scans the image data in search of people who are not wearing masks. 

The device could have been installed in the hallways of shopping malls to supervise if customers are wearing masks to protect themselves and those around them. If not, automatic notification to the security team would be launched. 

(Yes, we know it sounds quite dystopian.)

Drug development and discovery

The power of artificial intelligence and machine learning can be used to search for the cure for new diseases, including the COVID-19. The internet comes with frequent news on the AI being used to support scientists in search of new molecules or chemicals to solve existing problems. The search for new battery materials is one of the recent examples. 

In the same matter, the search for a cure for COVID-19 or any other disease can be supported by AI-based solutions. A good example comes from Google-owned Deepmind, where the neural network designed to predict the way of protein folding was retrained to support the search for new drugs fighting the coronavirus. 

The prospects are impressive and breathtaking sometimes. But one would feel concerned, especially when reading about the facial recognition technology used to spot if people are wearing masks. Thus, the key is to approach the problem with care. 

How to use the COVID-fighting AI responsibly

The AI-based solutions are shown not only as a way to change the world for the better - but also for the worse. That’s why companies and governments need to be extremely cautious when using this technology, especially dealing with outbreaks. 

But there are multiple ways to tackle the problem in an ethical way: 

Data anonymization

Not long after the GDPR madness both the companies and individuals are highly sensitive regarding personal data. The former are afraid of high penalties for being incompliant, while the latter gained a significant deal of knowledge and consciousness about their rights and privacy regarding personal data - there was a huge buzz after all. 

To make the solutions compliant with all the data-related regulations it is crucial to ensure the anonymization of the data - revealing any medical data regarding individuals would be a huge violation of the privacy-data regulations. 

Staying case-specific

Solutions fighting the COVID-related issues can interfere with numerous regulations or be quite disturbing regarding the matter - just think about the mask recognition mentioned above. It could be tempting for officials to abuse these solutions - and that’s why it would be crucial to limit the usage of the solution to the COVID-related cases only. 

Limiting the time

Another way to approach these problems in an ethical way is to limit the time of running the solution. Disabling the usage of the solution after solving the problems connected with the outbreak would be a good way to ensure the solutions after the pandemic will remain compliant. 

A good example - but not related to the AI - comes from the EU, where Vodafone, Deutsche Telekom, Orange, and five other telecoms providers have agreed to share mobile phone location data with the European Commission to track the spread of the coronavirus. 

The solution is both limited in time and case-specific, thus the risk of abusing the new regulations remains limited. 

Summary

To a person with a hammer, everything looks like a nail - and the same goes for AI companies and ML-based solutions. The key is to remember, that the AI is only but a tool - nothing more, nothing less. It will not fight the virus alone or deliver a panacea for all diseases. But it can significantly support the work of people in the front line and back-end, be that a doctor, a clerk supervising the processes, or researchers looking for a new drug. 

If you would like to get some new information about AI solutions and how healthcare can benefit from using this technology, contact us now - for sure we can build something great together. 

Oh, and by the way - stay home!