Internet, AR, genomic and robotic technologies could make a pandemic much less disasterous

These masks don't help the uninfected very much, except stopping you from touching your face.

Back in 2005 during the H5N1 flu, I outlined a number of potential technological solutions to dealing with urban shutdowns due to an epidemic. Sad to say, we don't seem to have done them. We may try some of them in the next (or current) epidemic but some of them needed some simple planning.

A summary of the plans included:

  • Plans for much more telecommuting and videoconferencing to facilitate working from home
  • Plans for video services (using spare TV channels and internet) to allow teachers to convene and teach class to students staying at home.
  • Plans to improve online shopping, including optimized grocery delivery for those sheltering at home, using delivery people with proper training, or who have received prioritized vaccines, or who are recovered victims with immunity.
  • In general, making the best use of recovered victims with immunity at all critical infrastructure.

At the time, the internet was not as widespread. I underestimated how quickly small amounts of vaccine can be generated -- it is possible to generate them, but takes many months to test them for efficacy and safety.

With a new potential pandemic underway, it's worth considering what is possible with new technology. In any event, a culture needs to plan for this, and do test runs to make sure it works, or there will be panic and chaos when you need to activate such plans.

New Options

  • We're really good at online ordering and delivery but may not be able to scale it up to everybody. As such there may need to be some "rationing" with a focus on getting essentials to people before wants.
  • Telepresence robots can be distributed more widely for places where video calls won't do. In addition people can be trained to work as a human telepresence robot if this makes sense. It should also be possible for the delivery logistics to deliver and pick up robots, and sanitize them.
  • Robotic delivery is now arriving. We have active delivery fleets in use at several campuses at Starship Technologies, for example, and several other companies are also working on it. These fleets are not ready for Covid-19, unfortunately, but they will be the next time.
  • Likewise, robocars are almost ready to move both cargo and infected persons without putting a driver at risk. Infected and quarantined persons can be moved with no other human needed. It may only be necessary to get the cars to deliver themselves and be driven by an infected person to get to medical care or quarantine. Vehicles can be designed for easy sanitation -- a good idea even outside of an epidemic. For example, if they can sit in a hot garage without damage.
  • Clearly the video conferencing and meeting tools are now up to the task outlined in my earlier article to convene schools and meetings easily.
  • DNA/RNA based infection testing can be made cheap and improved. See below for how to make it cheaper. Those declared free of the infection could, for a short time, use proof of that to fly or congregate with others so tested, if they maintain isolation after the test.

Cheap and frequent testing

Thanks to genomic advances, RNA based tests are available for many diseases. If there is a test which is highly sensitive, and only a small percentage of the population is infected, you can use a bulk testing approach to do low cost testing of large populations.

To do this, you test a group of people who are unlikely to be infected -- less than 25% chance in the whole group, and ideally much less. You gather multiple samples from the whole group and then mix together one sample from each subject. If the resulting mix is clean, all subjects are clean. If the mix tests positive, you divide the group into 2 groups and repeat the process, until you are down to a group of 4 whom you call in for individual re-tests if needed.

This can be even more efficient than pure binary search, because half the time you will get lucky when testing two halves, and get a positive on the first half. The second half is unlikely to be positive, but in case it is, you mix it with other unlikely groups to cheaply confirm all are negative.

This only works if your test has a very low false negative rate. The false positive rate can be slightly higher but also should be low. And the population being bulk tested must have a very low infection rate. Typically these would be asymptomatic people just wanting to be clear. Unfortunately the current test in use for Covid-19 is not very accurate and this approach would not work with it.

You also want to test people with symptoms cheaply. In many cases with Covid-19, symptoms are mild and people may not be sure what they have. If they learn conclusively that they have the virus, they can know to isolate and seek appropriate care, and when they recover, they can add themselves to the highly valued ranks of the immune, who will find ready work if they want it in providing infrastructure.


In China, the virus pushed the use of telemedicine for all non-virus doctor visits which could be easily done remotely. Telemedicine is not new, but it's had a hard time getting a foothold, due to medical rules insisting a doctor physically examine a patient, as well as insurance rules. But for many conditions, it is just fine, and aside from having no infection risk, also is quite a bit cheaper and faster for the doctor and patient. It is a key tool to lower health care costs. And an epidemic is a great reason to develop it more.

In the future, each home of any means should have a small diagnostic tool which can connect to a phone, and allow doctors and nurses to measure all vital signs, listen to the lungs and heart and even look at samples under a microscope or perform basic lab tests on urine, saliva and finger-sticks of blood. The majority of visits could then switch. Such devices could also be available for those who can't afford one to the home at various retail locations.

The immune and Augmented Reality

A lower skilled worker maintains a machine, guided by a remote specialist

There may be no vaccine, but those who recover from the disease will almost surely be immune. Thanks to RNA or antibody based tests, we can verify that they really had the disease and are now immune. Systems can be put in place to put them immediately to work. Of course, if their own field is essential and not shut down due to the crisis, they would do that, but plans can be in place to catalog the immune and offer them jobs of various sorts in essential infrastructure services. Many could receive remote video training in things like logistics and driving delivery trucks, and basic medical assistance.

In addition, we could equip many people with AR headsets with cameras, and allow them to work many jobs, supervised by the actual workers of those jobs remotely. This is called remote assistance. In effect, they would become very smart and capable robots. The remote worker would be able to tell them "now push that button" and to draw a circle around the button in their field of view. This sort of augmented reality technology is already being used for worker training and remote maintenance of complex equipment. It can also be used with skilled but not specialist workers -- for example a nurse might be guided by a remote doctor able to see out the nurses' headset and talk to the nurse and the patient.


Check out this early days article comparing the sensitivity of RT-PCR (genomic testing) with CT (chest x-rays) for COVID-19 diagnosis and screening. It shows CT has the best sensitivity, so far. It also makes the good technical distinction between sensitivity, specificity and accuracy. For containment, sensitivity is the important number.

Obviously CT can't be done in bulk. Surprised that CT would reveal that much though it would show the pneumonia or other inflammation in the lungs, but not name their cause. Anyway, to do cheap bulk testing we obviously need a highly sensitive DNA/RNA based test which we may not have. China is now manufacturing over 1 million tests per day, though.

Add new comment