BlueDot is the Canadian startup that made headlines after giving advance notice of the arrival of the pandemic virus now known as COVID-19. While on January 6, 2020 U.S. Centers for Disease Control and Prevention, and the World Health Organization on 9 January 2020, were reporting the spread of the virus, BlueDot sent out a warning to its customers on the morning of December 31, 2019. It should be noted that the same system correctly predicted which cities would be at first risk of the virus outbreak (Bangkok, Seoul, Taipei, and Tokyo).
BlueDot, among also other products, offers its customers access to the “Insights platform”, within the service, with the help of artificial intelligence, informs subscribers about emerging infectious diseases and the risks associated with them. The exact way the platform works remains a company secret. However, it is known that the platform analyzes data on airline ticket purchases, flight schedules, local public transport, posts on medical forums, blog entries (interestingly, posts on social media platforms such as Facebook are not analyzed because they are considered to be too chaotic), information on animal diseases and demographic data. In addition, BlueDot collects climate and temperature data in a given area (including using data collected by satellites). In some cases, which will be discussed later, the startup’s artificial intelligence is also fed with data on the movement of mobile phone devices. All of this data is then combined with official disease reports and health organization reports. The results of this work are then passed through the information already gathered about the symptoms of diseases, how they mutate and how they spread.
The platform processes this data in 65 languages. The system is not only trained directly to detect relevant data on the risk of spreading epidemic diseases but also learns natural language processing (NLP – natural language processing).
However, this system is not fully autonomous, it is quite significantly assisted by scientists on an ongoing basis. I’m also referring to a large group of doctors who both help train the machines and interpret their results.
The system of collecting and processing data operates permanently, which gives it a huge advantage over entities based on an analysis performed solely by humans. Another advantage of this tool is that it is not subject to emotions. While the world was focused on the COVID-19 pandemic, BlueDot’s algorithms picked up another dangerous virus spreading in West Africa called Lassa Fever.
The credibility and power of this platform led to official cooperation from both the Canadian government and a Canadian airline, and praise and thanks were publicly expressed by Prime Minister Justin Trudeau for helping to contain the outbreak.
These circumstances have influenced collaborations with the World Health Organization, the city of Chicago, the state of California, and even Facebook. It is estimated that this collaboration resulted in better coordination of activities, allowed for anticipation of health care occupancy – and thus hospital overcrowding. The collaboration made it possible to verify in real-time whether individual restrictions were causing reduced mobility and, in short, made it possible to verify that people were staying at home. Controversy arose over the locations of mobile phone devices provided by the city and state governments. According to statements made by the city and the state authorities – the data was anonymized.
This situation undoubtedly affects the disparity between the entrants to this type of service and BlueDot, which has trained its algorithms based on data that is not widely available. Once again, the question arises to what extent the results of such work are owned by the company. The very important issue of government making public health data also emerges in this context. I promise to look at this last issue soon.
The notion that artificial intelligence predicted the coming COVID-19 pandemic to this extent and with these consequences – is exaggerated. It is true, however, that it did predict the manner and rate of its spread. Of course, at the time, the reliability of such predictions was in question. How accurate, reliable, and therefore useful such analyses remain an open question. Contrary to skeptics, such reports cannot be considered merely a curiosity. As the BlueDot platform showed after the outbreak of the pandemic – these tools can be successfully used to limit the spread of the virus.
Ultimately, such tools will not only make it possible to correctly estimate the manner and speed of the spread of infectious diseases, but also to calculate the impact on social life and the consequences for the economy and political life. The price for such social security is a partial surrender of privacy and vulnerability to social engineering techniques.
Guest Author: Michał Pankiewicz
