Updated: May 6, 2020
Abandoned offices. Shuttered restaurants. Snaking queues in the supermarkets and pharmacies. Borders re-discovered. COVID-19 has managed to take us back to a previous century with public health officials doing their best to intervene in the relentless spread of this virus with nationwide restrictions.
But we should not overlook the enormous amount of data we could have - and need to balance societal good against individual freedoms. Two obvious observations to start. First, you can only catch COVID-19 from someone who is infected with the virus. Second, many of us are wedded to our smartphone, and rarely leave it from our sides. South Korea has been the pioneer in linking these two observations, and publishes online the movement history of infected individuals, allowing those that can remember where they were to check where "bumps" may have happened, and go to rapid drive-through testing centres. Other countries are moving in this direction. Many app developers are exploring what may be possible in bringing this information to the individual. But this is 2020 and the world of connected digital ecosystems. Maybe we can go a lot further than this and build an accurate picture of exposure and risk across our populations. Lessen the need for blanket bans that could isolate many for months with all the attendant impact on state of health and mind. The challenge at the moment even in South Korea is that we are dependent on individuals remembering where they were and taking action. Instead, imagine if we were able to use the information of where infected individuals had been to identify those that were likely to have been infected, and trace forward where they had been in the meantime and are now. This is no more than contact tracing, but the difference is that machine learning would enable us to carry out on a scale beyond the resources of any public health observatory. The algorithms that enable our smartphones to determine when we are walking as opposed to running would successively improve our estimations of likelihood of infection. Were they facing me? How long were we together? Did the smart phone detect a cough or a sneeze in that time, picking up on the sudden motion? And when the likelihood of infection breaks through a threshold, an app would direct you to self-isolate and seek testing, while directing others away from you in the same way that navigation apps steer cars away from traffic. Machine learning linking those negative and positive tests to improve still further the predictive accuracy. There are undoubtedly many questions to work through so that location data could be shared safely and anonymously without compromising the well-being of the individual. It may seem a step too far for many. But as we move into a new world with increasing blanket restrictions that fail to capture the actual risks that you face, the balance of opinion may shift. Our smartphones may be our best tool in the fight against COVID-19.