When Privacy Preserving Technologies are Decentralized they become Dignity Preserving Technologies
I had a vivid dream where I was taken against my will to the emergency field hospital at Javits Convention Center, after a positive contact tracing signal and no available tests turned me into a “health risk”. As I entered the makeshift facility with no valid immunity passport, soldiers handcuffed me with an electronic bracelet, that I couldn’t remove…
I woke up checking if a number wasn’t tattooed on my forearm…
Luckily the facility at Javits Center is closing, centralized government-sponsored contact tracing is seriously being reconsidered, and electronic wrist bands are used in Bahrain and not in New York. And although we still are not testing enough, we have a pretty good idea what needs to be done:
Mass testing, crowd intelligence and decentralized tracing.
#1: Mass Test. Identify. Isolate. Repeat.
We know we can not devise sound public health policies without widespread testing. With up to 50% asymptomatic carriers and endemic pockets in nursing homes, prisons and food processing facilities, it is impossible to imagine how and when to relax social distancing in a way that will avoid a second wave of infections. Unfortunately we are testing far too few (below).
For an aggressive test-and-isolate approach to work, an estimate of 5–20 million tests a day are necessary to reduce the number of people that need to be isolated, allowing the remainder to return to work. This in return will reduce the burden of tracing contacts, wearing masks and maintaining social distance. Economically speaking, 35 million tests per day at an annual cost of $100 billion, is a fraction of the $350 billion in monthly losses due to the ongoing lockdowns and social distancing measures in the US (for details read Paul Romer’s excellent report: Roadmap to responsibly reopen America).
#2: Create crowd intelligence
As of April 26, 12,067 people died from COVID-19 in NYC. However this probably is an underestimate, since as in many places, official daily figures exclude anybody who did not die in hospital or who did not test positive. With the cause of death taking several days to establish and not counting people dying from other conditions that otherwise would have been treated successfully, it is difficult to gauge the success of any public health policy.
Therefore it is not surprising that applying social distancing in the US for example, differs radically from states who apply strict social distancing (NY, NJ, CT, MA, CA, WA) to some (FL, GA, TX, TN) to no social distancing (SD, NE, OK). Furthermore, since predictive mortality modeling diverge, the use of a composite analytical framework might be a better guide to define the success of such policies.
One way to compare outcomes is to use the Deep Knowledge Group methodology which compiled data from over 200 countries and ranked 60 of them. Countries were evaluated using 24 specific parameters in 4 distinct categories: (1) Quarantine Efficiency, (2) Government Management Efficiency, (3) Monitoring and Detection, and (4) Emergency Treatment Readiness. The results can be seen below,
This type of context-rich health data has been a valuable market for at least the last decade. Clinical notes, lab and imaging results, genomic and wellness data added to insurance claims, purchasing and social media input has contributed to an already saturated 2.7 zettabyte (2.7 trillion gigabytes) digital universe.
However this universe, as COVID has shown, is fragmented, uncoordinated and to a large extent fragile. Health data might be designed for daily operations, but it is not organized for a multi-party crisis management, which requires real-time research and analytics. The presence of many intermediaries like enterprise data warehouses, data aggregators, administrators of patient and government registries (see below) have created an attack-, collusion- and censor-vulnerable environment.
Unfortunately, even with COVID, we continue to use these data aggregators like ArcGIS (Global), Anodot (US) and Covid Tracking (US) to provide up-to-date public information originating from government and public health agencies and research institutions, without leveraging the real-time capabilities of federated learning, combined with privacy preserving technologies (like ZKP, TEE and Homomorphic encryption) and blockchain. (I wrote about our work at Consensys Health here).
#3: Trace Ethically
We underestimate the already existing traceability technology available. Anywhere from mobility maps from Apple to heat maps from Tectonix Geo. We are all surveilled and it seems that many countries have already agreed to use contact tracing (below).
Whereas China and South Korea have applied or the UK wants to apply government-sponsored tracing (which includes the use of credit card transactions, smartphone location data, and CCTV video footage), other countries (including the US) are examining the use of contact tracing with transparent oversight, clear principles of fairness (including equal access and treatment), robust data protection, and audits of the algorithms used.
Still most Americans don’t think cellphone tracking will help limit COVID-19 (below) and are split (52% vs. 48%) whether it is acceptable for the government to use people’s cellphones to track the location of those who have tested positive for the virus to understand how the virus is spreading.
One of the efforts to protect personal privacy lead by MIT, is building an open, interoperable, privacy-preserving protocol called PACT (Private Automated Contact Tracing). PACT is designed to be a technical standard/specification that anyone can deploy on any smartphone without revealing private information to other individuals, the government, health care providers, or cell service providers.
Final Thoughts: mass testing or mass surveillance?
As lockdown and other restriction measures are progressively withdrawn, dozens of contact-tracing apps are being built for large scale adoption by many countries to avoid or minimize contagion resurgence. However a centralized approach, where data is sensed by the app and sent to a nation-wide server, raises serious concerns about citizens’ privacy.
As written by Mirco Nanni et al. “…We advocate the conceptual advantage of a decentralized approach, where both contact and location data are collected exclusively in individual citizens’ “personal data stores”, to be shared voluntarily, only when the citizen has tested positive for COVID-19, and with a privacy preserving level of granularity…”
But there is something more important than preserving privacy using a decentralized compute architecture. It is to preserve our dignity because…
… we are our actions and our actions are who we are. These actions are captured by data and every time these data are used, abused or sold by a third party, a part of our dignity is stripped away. These technologies are not merely privacy-preserving, but are dignity-preserving.
Unfortunately history is replete with examples of dignity-stripping chapters. COVID-19 has provided us an opportunity, with blockchain, to keep us healthy and safe without loosing our dignity.
If you liked what you read, go ahead and “Clap” below so others will see it too (up to 50 claps allowed!)
COVID-19 Data Is Valuable Because We Are Valuable was originally published in Data Driven Investor on Medium, where people are continuing the conversation by highlighting and responding to this story.