AI is Mapping the Coronavirus and Inferring its Possible Economic Impact
AI has proven successful in helping to monitor the outbreak of the coronavirus (2019-nCov).
An AI-driven company, BlueDot, detected the outbreak on December 31st, while the World Health Organization only warned the public on January 9th. BlueDot is using natural language processing and machine learning to scrape 100,000 daily news articles written in 65 languages and track any outbreak concerning over 100 infectious diseases.
BlueDot is also using a wide variety of information such as flight path records, livestock disease records, and even plant disease records to better anticipate the evolution of any outbreak.
Metabiota, a competing AI-driven company to BlueDot, is working with the same tools and successfully made correct predictions regarding the spread of the coronavirus to areas such as Hong Kong about one week before the authorities detected new cases in the area.
BlueDot declared they will not use social media because it is deemed too noisy, however Metabiota claimed it used social media forums to obtain more clues, i.e no consensus exists on the use of social media as a reliable source of information.
These companies still do not entirely rely on AI. The firms find it necessary that experts in the field of interest be involved at the end of the process to check the relevance of the predictions. At BlueDot, epidemiologists are always at the end of the chain as safeguards of the scientific relevance of the AI predictions.
Currently, BlueDot sells its reports to government health officials, airlines and hospitals, while Metabiota increasingly focuses on private clients. Metabiota is trying to go further, in not only predicting the spread of the coronavirus, but also analyzing the potential effects of such a pandemic on the economy.
Metabiota has the idea that “we can’t eliminate epidemics but we can minimize their impact”. Metabiota is especially targeting the insurance and reinsurance industry by offering to model the potential losses of a company based on different scenarios. In that way, the insurance industry can better quantify their exposure and manage their earnings in the event of an outbreak, as well as creating a new product to suit pandemic-related events.
In particular, Metabiota gives the example of South Korea, where the tourism industry suffered a 41% decrease in activity because of the 2015 MERS outbreak, estimating the loss at over $90 million. Now that the world is increasingly connected, epidemic outbreaks are to be more frequent and might become more massive.
If the coronavirus epidemic is to continue expanding, the economic impact could become immense. When a city as populated as Wuhan (11 million inhabitants) is put under quarantine, consumer spending stops, which reverberate negatively in all aspects of the economy. Therefore, through natural language processing, Ai scrapes information and predicts the future spread of epidemics; and through machine learning levying a proprietary data set of former epidemics, used to evaluate the potential economic impact of epidemics, AI has become a key to mitigate the aftermath of such epidemics.
Written by Charles Ballario & Edited by Alexander Fleiss
https://venturebeat.com/2020/01/31/ai-weekly-disease-coronavirus-prediction-spread/ https://www.wired.com/story/ai-epidemiologist-wuhan-public-health-warnings/ https://bluedot.global http://www.metabiota.com http://metabiota.com/sites/default/files/presentation_files/metabiota_brochure_sept_201 7.pdf http://metabiota.com/sites/default/files/presentation_files/Metabiota%20Pathogen%20Sen timent%20Index%20Flyer_Apr%202018.pdf