A Geiger-Counter for Viral Risk

Steve Hazel
6 min readMar 13, 2020

--

A killer app for Google Glass and things like it?

Geiger-counters tell us how much dangerous ionizing radiation is in the area. Personal dosimeters do much the same thing, measuring how much radiation we’re exposed to while on the job. With these measurement tools, we learn to avoid the invisible yet deadly risks of aggressive cancer, radiation burns, and radiation poisoning.

I propose we kick off a virtual Manhattan project to build a “viral dosimeter” along with its underlying connective system. This “viral dosimeter” measures and exposes the largely hidden yet deadly risks of contagious viruses like COVID-19, H1N1, SARS, Ebola, and most importantly all the new viruses we can expect to surface in the future.

By combining facial recognition, augmented reality, geo-location, and cloud infrastructure with wearable cameras (such as our Smartphones), we have the ingredients at hand to quickly create a variety of prototypes of personal “viral dosimeters” that measure an individual’s infectious risk, whether in times of global pandemic or during the months and years between pandemics.

By building and deploying such a tool, we’ll soon see who is serious about multiplicative risk and who is not.

Quantifying Your Viral Infection Risk

The way I see it, your viral infection risk score (VI) is the combination of two separately-measured risk scores: your social distance risk score (SD) and your viral vulnerability risk score (VV).

VI = SD × VV

Do you know what your social distance score is? Do you know how it compares to the people you met today? Probably not. You might want to know though, because your life could depend on it.

Think of your social distance score as a speedometer showing how rapidly a virus might reach your doorstep.

Do you know what your viral vulnerability score is? Do you know how it compares to your friends and family? Not likely, but your life could depend on knowing. Their lives could depend on it too.

Think of your viral vulnerability score as the speedometer showing how rapidly you can expect to be overcome by a virus that makes it to your doorstep.

To get your viral infection risk score, multiply your social distance score SD, which ideally is a small number, by your viral vulnerability score VV, which is also ideally a small number.

Once again:

VI = SD × VV

Basically, a virus won’t get far in a population of people with low VI scores, but it will spread like wildfire in a population with high VI scores.

Let’s see how we can calculate these scores on the individual level.

Social distance score

Your social distance score shows how likely you are to be exposed to a virus that is spreading out in the world.

For a system to calculate your basic social distance score, it would have to make a record of each unique person you’re in contact with, how often, how distant, and for how long. Then, for a full score, it would fold in the social distance scores of the people you encountered, provided the data is available. The farther in the past an encounter, the lower its weight. Perhaps a 30-day rolling window is enough.

An always-on camera with facial recognition can achieve this because it enables an accurate, real-time count of the people you’re around with no intervention necessary. When the people around you are also measuring like this and sharing their own real-time risk score, everyone gains clarity into the overall situation.

Remember Google Glass? It didn’t have a killer application, but this could be it. Imagine looking around you and then being warned in real-time about the specific people (and general situations) that represent serious viral threats.

Keep in mind though that a central database of faces isn’t required; the most important thing is an accurate personal count, not full-on identification and tracking.

Geolocation becomes quite helpful when it comes to filling in the gaps. When there’s geo-evidence that you were in the same place at the same time as a high-risk person who you can’t otherwise identify, your own risk score would rise for a time.

To reduce your score, simply raise your “self-quarantine” flag for a while. You might enjoy the health benefits of doing this on a regular basis, only now nobody is going to hold it against you.

Viral vulnerability score

Your viral vulnerability score shows how rapidly you can expect to be infected and waylaid by a virus you become exposed to.

To calculate your viral vulnerability score, the measurement tool would have to first account for your age, current health condition, and sleep quality. As well as any other vulnerability vectors we know of or discover later.

From there, it could use the always-on camera to count how many times you touch your face, touch other people, and touch various surfaces or objects. It could measure how much time you spent outdoors (away from people) and perhaps even how much time you spent sleeping, among other things. Presumably, the measurements would get better over time. Measurements from others sources, such as outdoor temperature, indoor temperature, body temperature, and air quality could also be folded in.

Modern AI machine learning techniques, such as the deep neural networks that allow self-driving cars to recognize what’s happening in their environment, are quite capable of this sort of pattern recognition. While in a self-driving car the real-time consequences are life-and-death, in the case of a viral dosimeter we care mostly about roughly measuring behavior patterns that we can then gradually adjust. 100% accuracy is not necessary.

Like with the social distance score, a 30-day rolling window of relevance would make sense, while keeping the option to retain a detailed historical record. There’s no technical reason why any raw data would have to be retained by third-parties, but likely there are strong economic reasons. We’ll have to see how the market shakes things out.

To reduce your viral vulnerability score, there’s plenty that you can do:

  • Be as fit and healthy as possible
  • Get plenty of solid sleep, and do whatever it takes to get it.
  • Eat healthy
  • Exercise often
  • Get outside for hours each day
  • Touch your face less often.
  • Carefully wash your hands more often.
  • Disinfect key area of your home or office more often (or less often if it’s too sterile!)
  • Filter your air
  • Take your medication
  • Associate with people who have VI scores similar to your own
  • Measure your local air quality

And so on. To increase your viral vulnerability score, simply do the opposite.

In the case of infection

If your lifestyle or career happens to be relatively high risk, you’d be wise to prepare for quarantines. Same if you live in a region that is generally high-risk due to population density, low air quality, or cultural habits.

If you become infected, you’d either be flagged by a medical professional or else flag yourself, and this flag would be shown beside your score (perhaps along with a history).

If you neglected to be flagged despite an infection, the rest of us will have a much easier time tracking you down and enforcing quarantine and/or guillotine.

There’s no better time than now

Overall, this “viral dosimeter” that we could quickly cobble together from already-existing technology enables us to see the risks we’re facing both individually and collectively. With it, we can learn to avoid the invisible yet deadly risks of life-threatening viruses.

Is it wise for us to cluster our most vulnerable into nursing homes, hospitals, cruise ships, and trains? Perhaps we’ll learn a little.

Is it wise for us to go to work or school while infected? What is the real consequences? Perhaps we’ll learn a little.

We have a lot to learn and much to do. Not coincidentally, in the coming weeks and months millions of bright, capable, and worried people are going to have a lot of time on their hands. I’ve seen many people ask what’s good to do during this time of self-quarantine and social distancing. This viral dosimeter could be what we’re looking for.

How many virtual hackathons do you think it will take for the first prototype to hit the streets? How many prototypes will it take for one to stick?

--

--