Not Immune


The human immune system is a wonderful thing. In an environment where there are thousands of types of cells doing a wide variety of specialized things, it’s generally able to detect when something out of the ordinary happens, run through a variety of possible responses, adapt to the results and craft an effective counter that restores normal conditions quickly and with acceptable (even minimal) disruption to the system as a whole. If it’s seen the disturbance in normal operations before, the reaction is even faster, with specific counter-measures rapidly adapted from generic “weaponry”. As we have come to understand how the system works (an understanding that’s still far from complete), we’ve been able to improve on the natural system’s abilities by sensitizing it with vaccines that help to build defensive patterns without exposing the whole system to a potentially deadly threat.

It’s a wonderful example of how an adaptive approach can work within a complex system that can’t be fully defined a priori, and like a lot of wonderful, but context specific, examples, we have a tendency to generalize the model to other areas we’d like to protect from seemingly similar threat patterns. That’s not always a good idea, especially when we push the analogies further than we should – and then ignore the shortcomings.

First, the immune system isn’t perfect, falling short of perfection in several ways. The very mechanisms it employs to create and deploy defenses can run out of control, with the weaponry destroying healthy cells as well as intruders and causing the entire system to collapse. Second, attacks on the base mechanism of the immune system itself threaten the whole system as every defensive capability is compromised. When this happens, even routine issues that would not normally be a problem become potentially lethal. Third, the analog of a vaccine – a major force multiplier at the population level in humans, doesn’t always translate well to other contexts. Fourth, diseases evolve too, with some even mimicking “normal behavior” until they can overwhelm the immune response.

But perhaps most importantly, the immune system model requires that some people get sick, even die, so that most the population can evolve effective defenses. Until we see disease mechanisms in action, we can’t design a vaccine – and until the immune system sees a new threat, it can’t specialize its generic weaponry to counter it. Sometimes we never get a workable defense – usually because the disease is so lethal that it kills the local subpopulation before it can spread very far. That’s a natural “containment” mechanism – and an aspect of the immune system as a whole (population level rather than individual) that’s often ignored.

Despite these shortcomings, immune system models can teach us some important lessons:

  • Understand what “normal behavior” looks like, so that it’s easier to detect anomalies
  • Classify anomalies, so that their potential impact can be evaluated and an appropriate response defined
  • Remember past anomalies so that response times can be reduced for subsequent “infections”
  • Beware of incidents that mimic “normal” but seem out of context
  • Think about how containment works in a system that cannot be perfect at prevention
  • Share information about effective countermeasures so that others can react faster and more effectively
  • Test the system regularly to make sure it’s detecting unusual behavior (a workable analog for a vaccine)

Do all these things (it’s not supposed to be an exhaustive list – you can probably think of others) and the immune system model can be effectively applicable in other threat-rich contexts, such as supply chain or cyber security. Assume the biological model transfers completely and exactly the same issues that confront the immune system in people will show up.

Vaccines are great, but prevention and hygiene matter too. Be well.

John Parkinson
Affiliate Partner
Waterstone Management Group