A special tool named “BlueDot” has informed its customers on December 31 that the possibility of a dangerous virus outbreak is very high. This warning came nine days before the World Health Organization notified the public about the novel coronavirus outbreak in China, and six days before the US Center for Disease Control and Prevention put out a warning about the same. But how exactly has the AI-based Canadian health monitoring platform managed to achieve this feat?
BlueDot employs complex algorithms that monitor a large number of news reports from around the world. The sources include animal and plant disease outlets and official proclamations on 65 languages, including Chinese. BlueDot can even pick up forum reports from individual users, or comments of people under blog posts. The only exclusion from the sources is social media posts because these make the situation too chaotic. The key in BlueDot’s data sources though is in having access to global airline ticketing data. This helps the AI predict how a potential virus could spread, and what infection rate patterns apply.
In the case of the coronavirus, BlueDot predicted that it would jump from Wuhan to Bangkok, Tokyo, Seoul, and Taipei in just a couple of days, which it did. If the authorities had paid attention to BlueDot’s reports at that time, hundreds of people could have been saved, thousands of infections could have been prevented, and the speed of the outbreak could have been reduced. These reports are sent to the public health officials in twelve countries, as well as a selection of airlines and hospitals free of charge.
The creator of the epidemiologist AI, Kamran Khan, tested several predictive algorithms before ending up with the set that is deployed in BlueDot back in 2014. When the system was ready and presented to prospective investors, the project raised $9.4 million in the first round of capital funding. In 2016, BlueDot proved its worth once again by predicting the international spread of the Zika virus from Brazil, in 2018.
