AI and its unmatched ability to analyze data could be used to predict the next deadly virus outbreak. The revelation comes by way of ProPublica, which found that artificial intelligence trained on Ebola-related environmental data was able to flag outbreak areas. The technology can also highlight future outbreak sites that are often overlooked by government agencies.
The process involved training the AI to analyze past outbreaks, identify recurring patterns, and subsequently recognize areas with similar potential risks.
For months, a team of researchers dedicated their efforts to educating a computer system about Ebola. The AI was provided with extensive information regarding the geographical and demographic aspects of regions where the virus had historically surfaced. The process involved training the AI to analyze past outbreaks, identify recurring patterns, and subsequently recognize areas with similar potential risks.
Some areas flagged for the deadly virus were expected due to its historical impact on specific countries. However, the unexpected highlight was Nigeria, Africa’s most densely populated country. Despite never being the source of an Ebola outbreak, this West African nation, serving as a significant international travel hub, played an unexpected role a year ago.
Nigeria facilitated the transmission of another virus, mpox (formerly known as monkeypox), to Europe and the Americas, leading to a global spread. Although mpox is generally non-lethal, its ability to disseminate so widely was surprising. Due to their findings, scientists believe that health officials have disease control backward.
They argue that current practices are overly reliant on virus data after an outbreak occurs or from past outbreaks. “Being proactive is the best line of defense,” Solomon Chieloka Okoli, a Nigerian Field Epidemiology and Laboratory Training Network researcher said. “If you wait, a lot of people will have died before you can get yourself together.”
AI’s ability to study data on deadly viruses like Ebola could allow scientists to detect viral outbreaks early on.
Areas in nature where a virus initially transfers from infected animals to humans, referred to as “spillover areas,” require increased attention. Such locations often include forests that have undergone partial logging, creating open spaces where animals and humans can come into contact. Scientists worked alongside AI that was provided with data from regions inhabited by known carriers of Ebola.
The AI tool does not inherently consider environmental factors. This implies that health officials might have inadvertently created a deceptive and potentially hazardous assessment.
This dataset included information about deforestation and shifts in population. Subsequently, the AI algorithm assessed the significance of these factors in relation to past Ebola virus outbreaks. It then extrapolated this knowledge to villages situated in regions that have the potential to host animals carrying Ebola but haven’t previously experienced spillover events.
The AI flagged 51 locations with tree loss patterns similar to past Ebola outbreaks, with 27 in Nigeria. However, due to the rarity of outbreaks in the African region, Nigeria’s disease control agency probably won’t give the virus much priority. Moreover, any concerns it might have harbored were alleviated through an analysis conducted using a widely-used tool developed by the US CDC.
This tool examined recent virus outbreaks and concluded that Ebola held a relatively low position in terms of priority. As it happens, the tool does not inherently consider environmental factors. This implies that health officials might have inadvertently created a deceptive and potentially hazardous assessment.
Ensuring the prevention of a future virus outbreak shouldn’t be overly challenging, especially with the help of artificial intelligence. However, navigating the politics around technology will be difficult.