The Coronavirus pandemic has brought about disruptive technology solutions to slow the spread of the virus and minimize its impact. Contact tracing utilizing mobile technology is one promising but controversial solution that has been tested in some countries and proposed in the United States.
The Washington Post reports that millions of people around the globe have already been placed under some form of surveillance in an effort to monitor people’s movements and attempt to trace the spread of COVID-19.
Of particular note, at least 27 countries have already started using data from cell phones to track the movements of their citizens. While this approach has a great potential to benefit the public good, the implementation and adoption of such technology raise important questions about transparency, AI ethics, and data privacy.
Contact tracing can be one of the most effective ways to contain an outbreak. However, COVID-19 is not a typical outbreak. The virus is often transmitted by individuals who display no symptoms of the infection and may not even realize that they are carrying the virus.
Standard contact tracing usually involves individuals who are symptomatic and are aware they are carrying the infection in question. Because of this, traditional contact tracing methods are challenging and problematic.
The Big Data Institute at Oxford University has proposed a solution for a mobile contact tracing application that is much more agile, efficient, and scalable than traditional manual contact tracing methods. The team developed a mathematical model designed to stop the epidemic if implemented on mobile devices by a substantial portion of the population. It’s poised to reduce manual contact tracing from 72 hours to four hours. By replacing weeks worth of manual work performing contact tracing, the mobile application can slow the spread much more quickly than traditional methods.
First, the majority of the population – including symptomatic and asymptomatic individuals – must voluntarily opt-in and adopt the application. The research team at Oxford estimates that about 60% of the population would have to utilize the application for mobile tracing to reach enough new virus cases to make an impact on the spread of the virus.
Another challenge lies in widespread and accessible COVID-19 testing. For the application to work, the majority of the population would need to have access to reliable testing in order for the application to properly and fully survey potential outbreaks.
There is much evidence suggesting that the application of rigorous wide-scale testing coupled with the application of mobile technology can blunt COVID-19 infection and mortality rates.
A recent Stanford report demonstrates how Taiwan – a country just 130 km from the epicenter of the outbreak in China – was able to contain the outbreak without the draconian lockdown measures in place throughout many of the advanced economies.
How did Taiwan manage to limit and contain the spread of the virus? The report highlights five interconnected factors: pandemic readiness, national electronic health records database, wide-scale testing, big data analytics, and the use of mobile technology to track movements of individuals who tested positive for COVID-19.
Within 72 hours of the outbreak, a comprehensive case identification protocol was instituted based on travel histories. Individuals determined to be high risk were monitored through their mobile devices.
Health authorities were then able to trace the movements of high-risk subjects and mitigate the risk of further transmission through targeted isolation measures. Taiwan may serve as a template for how to mitigate future pandemics.
Innovative initiatives are underway by major tech companies, notably Google, Apple, and Facebook, to track and analyze how the virus is spreading and to gauge the effectiveness of social distancing measures.
Facebook’s Data for Good project is designed to track the movements of users to measure and anticipate potential outbreaks.
In the context of the COVID-19 outbreak, researchers, non-profits and public agencies can leverage the data – which is anonymized and aggregated – to evaluate and implement strategies to slow the spread. However, such initiatives raise concerns about transparency and data privacy rights.
There’s a strong case to be made that the implementation of mobile tracking applications represents an unwelcome intrusion to privacy rights.
The GDPR includes specific guidelines for the use of data in the context of epidemics. Recital 46 of GDPR stipulates that the collection and processing of personally identifiable information without data subject consent are acceptable when it is necessary for humanitarian purposes. The information includes monitoring the spread of epidemics and addressing humanitarian emergencies, such as natural or man-made disasters.
It is the responsibility of technology leaders and policymakers to implement ethically sound, transparent, and fair applications related to the use of AI-driven profiling technologies and the transmission of highly sensitive health information.
For mobile tracing applications to be widely adopted and therefore effective in slowing the spread of COVID-19, the first and most important step is the grain the public’s trust.
Principles of fairness and transparency related to artificial intelligence are re-enforced by the EU Commission on The Ethics Guidelines for Trustworthy Artificial Intelligence. Every organization should be committed to these seven fundamental guiding principles for the application of AI technologies:
These principles were further re-enforced by AI Now Institute in its 2018 report in which the organization asserted the importance of the public’s right to known which technology systems are impacting their everyday lives and how they are doing so. To that end, organizations must be transparent of the algorithmic systems used and their purpose, applications, and potential public impact.
Furthermore, while the use of technologies to help mitigate the impact of COVID-19 may be essential to protect public health and safety, they should meet the following criteria as proposed by the Electronic Frontier Foundation:
In order for these solutions to fulfill their intended objective and have the greatest positive impact, transparency and fairness are needed.
Before the public adopts new technologies at scale, they must trust these new and emerging innovations. The ethical use of AI should be evaluated not only on a legal basis but on a moral one as well.
It’s the only way for novel solutions to gain the people’s trust, inspire widespread adoption, and make the greatest impact – in times of crisis and in the post-COVID-19 world.