Covid Tracing Apps and Data

While the pandemic is spreading, many countries are deploying mobile applications which track users' location continuously.

Then if a person is diagnosed positive, the app will get a list of all people who that person has been in contact with for past few days, and alert them. This seems like a pretty good way to help us detect potential cases early. Let us see if there are any problems with doing this.

The value of "Data"

Computerized information that can easily be processed is highly valuable for analysis. We can easily visualize them using graphs and charts. If they are GPS co-ordinates, we can overlay it on top of maps. We can do calculations on them, find averages, general tendencies. Using advanced statistical methods we can up to an extent make useful predictions about future.

It is undisputed that data is an extremely valuable tool. But when it becomes problematic is, when that data contains personally identifiable information.

We will see what are the dangers of letting anyone collect personally identifiable information, ie. private information in a later section.

What data do Covid tracing apps collect

There are two types of Covid tracing apps that are being deployed world wide.

Type 1

Most of the Covid tracing apps that are being deployed works as follows:

It will ask your personal information, like Name, Address and Phone number.

Then the app will continuously monitor your GPS coordinates. It will also collect a list of nearby devices using Bluetooth scanning and sometimes WiFi scanning.

Then it will constantly upload these data to a central server.

This system is simple and straightforward. So simple that one or two undergrad students can sit together and implement one from scratch in a few days. I believe it is why this system is widely used.

Type 2

Since the pandemic started, many academicians, professors, engineers and scientists working in the Information Security and related fields, figured that we will need a secure, privacy respecting solution to help trace potential spreads of the pandemic. They also understood the huge risks associated with type 1 apps. So they came together from various universities and collaborated to design a better solution.

The resulting work is called DP3T - Decentralized Privacy-Preserving Proximity Tracing.

The design, and implementations are open sourced and provided at:

Design Documents - https://github.com/DP-3T/documents

Apps and SDK - https://github.com/DP-3T

How does it work:

The app generates a random ID for each user, and transmits it through Bluetooth Low Energy (BLE) to all nearby devices.

Then the app scans using BLE the nearby devices, and stores the list of all the random IDs it has detected, locally in the device.

When a person is tested positive for covid, his random ID is sent to the server, which then all the other users receive. Then the app will check if that ID is in it's locally stored list of contacted IDs, and alert the user if needed. It is the responsibility of the user to then go to the health center to get tested and the process continues.

This is only a very simplified overview. Please see the DP3T white paper to see the additional security measures taken to ensure the users' privacy.

Independent of this, Apple and Google has collaborated to design a very similar system: https://www.apple.com/newsroom/2020/04/apple-and-google-partner-on-covid-19-contact-tracing-technology/

A brief comparison between DP3T and Apple-Google solution and other similar systems is given in the DP3T repo: https://github.com/DP-3T/documents/blob/master/README.md

The two sides of "Data"

As we saw, data will be an invaluably helpful tool during this pandemic. The ability to collect and process this much amount of data is only a recent development. We all know the dangers of leaving a shop open unattended in the midst of a crowded market. Most of us are familiar with the dangers of sharing OTP and CVV numbers to strangers. But we are not very familiar with the dangers of having huge amounts of data about us, including very private information, in the hands of strangers.

When automobiles first started appearing, there were not many rules for driving. For those vehicles were very slow, barely able to harm anyone. Only when faster automobiles were developed and they injured and even killed quite some number of people, we started to think about traffic rules.

We are in that period now. Recent developments have made data easier to collect and analyze in huge amounts. But we are still in the process of figuring out how to make data safe.

We already witnessed a few incidents in the past two years where bad actors misused the data collected by tech companies in alarming ways.  

See Cambridge Analytica scandal - abuse of private data for political gains: https://en.wikipedia.org/wiki/Facebook%E2%80%93Cambridge_Analytica_data_scandal

See Polar fitness app data breach of military personnel: https://www.washingtonpost.com/news/worldviews/wp/2018/07/18/fitness-app-polar-revealed-not-only-where-u-s-military-personnel-worked-but-where-they-lived/

Is the data collected by Covid tracing apps that sensitive?

Let us consider Type 1 apps.

Most of us are stuck at home anyways. So what's the worst that could happen if someone else have our location?

We will stop using the app soon after the pandemic. So even if the data breaches in the future, it will only contain information about where we were during the lock down. Which will mostly be the location of our houses.

Even if something could happen, what is the probability that it will happen?

Is it enough to weigh against the value of tracing solutions to health workers?

To answer these questions, let us have a closer look at the information collected by type 1 covid tracing apps.

They generally consist of name, address, phone number.

They also contain GPS location of the person once every few minutes. If we take that and sort it by the amount of time spent at a place, the top one will be their home. So we have the GPS location of their home.

They also contain the list of all people that person interacts with.

 If somehow you got this data, what is the worst thing that you can do?

Now if you are a normal person like me, who doesn't have a grudge against the society, nothing much. But imagine you are a terrorist, a terrorist who believes that everyone of a particular religion must be killed. What will you now do with this data? We can infer the religion of a person with reasonable accuracy from their name. So what you have now is the GPS location of homes of every citizen of a country sorted by religion. Any semi decent nerd can turn this into a mobile app that overlay these locations on google maps. Imagine this in the hands of a group of terrorists, and they sneak into your home town. Imagine what could happen.

But what is the probability that someone can hack and obtain our data?

It has happened already in Oakland, USA: https://www.clickondetroit.com/news/local/2020/04/16/map-containing-internal-oakland-county-covid-19-data-mistakenly-leaked/

Or suppose you are an authoritarian ruler. And there is this activist who is voicing against human rights violations in your country. This is becoming a headache for you both nationally and internationally. With this data in your hand, you have the list of all people who interacted with that activist during lock down. Sort this list by frequency, that is by number of times they were together. Now you got the name, address and GPS location oh homes of people closest to the activist. Perfect to exert some "pressure" on the activist.

What about type 2 apps? How safe are they?

They are designed around the mindset that data is like nuclear material. It is not safe to have them concentrated at one location. It does not upload your details to the server until you are diagnosed for covid. Even after uploading, the server will only contain the information that a random person at this location has covid. It doesn't store any of your private information. Even if a terrorist breaks into the data center and holds everyone hostage, they won't be able to get nothing that undermines the safety of our citizen.

Tracing data is invaluable for health workers

The usefulness of the tracing data is unquestionable. They enable health workers to rapidly act on points of spread and efficiently contain the pandemic. And we cannot afford to take such a useful tool away from the warriors who are fighting at the forefront for us. But here we have two options, both equally useful for health workers, but one more useful for bad actors.

Even though the times are difficult, we should discuss and know this issue deeper.

Let us be a responsible citizen and continue to do our part in the democracy we are in. Let us be grateful for our privilege to discuss and debate. The privilege to be part of a democracy, which a large share of the world's population still doesn't have.

And let us salute the doctors, nurses, health workers, government officials and politicians who are fighting at the forefront and working round the clock to save our lives.

--- airbender752

Airbender 752's GPG Public Key: https://rentry.co/tyy5u/raw

GPG clear signed version of this article: https://rentry.co/iya5h/raw