Novel coronavirus has impacted almost every country in the world. Millions have been infected, while fatalities caused by Covid-19 recently exceeded 100 thousand. Governments, science institutions and private companies are, among other methods, using Artificial Intelligence to find optimal ways of tackling the pandemic.
Searching for the cure
Using AI for finding novel drugs is not new. For several years now, large pharmaceutical companies have been collaborating with start-ups and established vendors for faster and optimal search for the desired medicine. AI has even been used to find new phase matter, which is solid and liquid at the same time. How about that!?
Well, it is not surprising that various research institutions and private companies have embarked on the journey of finding the cure for the Covid-19. First step in this direction was to understand the protein structure of the novel virus, which has been accomplished. AI played a significant role in this effort by proposing one of the structures, later confirmed by researchers from University of Texas. This is highly relevant for finding the vaccine and achieving preventing care of millions.
Next step would be identifying active substances which can fight the virus. Thousands of research articles have been written about the novel coronavirus and tens of thousands exist which are related to previous outbreaks and other relevant studies. AI is helping in parsing through the existing literature, searching for similar elements and relevant correlations to Covid-19 virus.
Forecasting the end of pandemic
Difficult but necessary social distancing restrictions have been applied in many countries. They differ from country to country and even within a country. These social distancing efforts and partial lockdowns have heavily impacted the economic situation in the entire world. Companies are losing money and uncertainty jeopardizes business continuity.
Questions about easing the proposed measures are arising more and more frequently. But how do we know, when the pandemic will end and can AI guide us to know the answer to this question? Or even better, could we have known that the pandemic was approaching?
One might say that just common sense would have been enough to predict the answer to the second question, but nonetheless there were several AI tools, like Bluedot which predicted the outbreak as early as December 31, 2019. We might have not listened to such alarming predictions previously, but we might learn to change our behaviour and pay attention in the future.
Answer to the first question seems to be more complicated, with many different factors in different countries contributing to complexity of the problem. Earliest example of the outbreak is, of course in China where studies as an alternative to epidemiological models have been performed to determine the length, size and the end of the outbreak and the results of such studies seem to be successful. The studies predicted the end of epidemic in China by the end of April. But predicting the same a global scale, is still challenging. The degree of uncertainty is very high, and many unknown factors might impact the predictions.
Responsible use of AI
In such difficult and extraordinary situation, we see some governments declaring national emergencies and obtaining more powers than usual. Unsurprisingly, AI has already been deployed by many governments to help them in slowing the pandemic. But we all know what AI needs. It is data and lots of it. To obtain such data, governments are using facial recognition to identify people breaking the quarantine measures and computer vision to control the social distancing guidelines on the streets. Furthermore, tracking people’s whereabouts using apps, fitness trackers and telecom providers, to accumulate the travel details is not only being discussed, but already implemented in many cases.
While this might be an effective way to stop the spread of pandemic during these challenging times, responsible use of AI must remain a priority and the return to normal privacy measures must be guaranteed. Lack of explainability and robustness, in addition to being often biased are the traits of AI which do not disappear because of the desperate situation the world finds itself.
Read more about AI & Healthcare in the blog published by Paul Fisher and learn more about the Covid-19 pandemic and privacy in the blogpost by Matthias Reinwarth.
Learning more about AI
KuppingerCole has an increasing body of research on the impact of AI will have on other sectors and integration with legacy architecture.
Reports include the following:
- Leadership Brief on Explainable AI;
- Assessing the Maturity of Core AI disciplines;
- AI in the Legal Industry and many more.
You can also contact one of our expert analysts for more specific information on an AI application or trend.