What Leaders Can Learn from Artificial Birds

With just three simple rules, a group of relatively disconnected individuals can form dynamic and useful group behavior.

How do you make birds flock in pursuit of prey and in avoidance of predators? And how do teams within an organization help the organization adapt to change?

Flocking is a complex group behavior. Birds aren’t taught how to flock and there is no centralized command structure telling them where and how to fly together. Yet, birds which have had no past experience with one another can exhibit complex patterns of collaboration and communication. Even more astounding, each flock is constantly inventing new patterns of flight in response to a changing environment, rather than following a predetermined course.

A researcher studying complex systems (such as an interacting flock of birds) often uses computer models, or simulations, to attempt to re-produce events similar to real-world phenomena – the more accurate the simulation, in theory, the more one understands about the full system. In a demonstration of this approach, artificial intelligence researcher Craig Reynolds created his program ‘Boids’ in 1987 as an attempt to understand the flocking behavior of birds. An early artificial life program, Boids was a virtual environment populated by a large collection of autonomous representations of birds which moved in space (‘boids’), and an assortment of obstacles such as walls to be avoided.

Reynolds couldn’t program his boids to flock outright. That would be cheating since there were no top-down roles or global rules like that in nature. Instead, he had to rely on local rules – rules which could govern the individual behavior of each boid – and hope that these rules, individually applied, could together bring about flocking. Moreover, Reynolds couldn’t overburden his boids with rules, these were dynamic and rapid behaviors in nature. He needed a set of simple rules for each boid to follow that could cause the entire population to move together in harmony, in real-time. Lastly, Reynolds couldn’t just program in a pre-determined pattern for the boids to follow, they must be able to adapt to novelty within their system (new obstacles and of course, moving boids).

After many failed attempts, Reynolds discovered that his boids needed only 3 simple rules to flock:

  1. Separation: steer to avoid crowding local boids and objects
  2. Alignment: steer towards the average heading of local boids
  3. Cohesion: steer to move toward the average position (center of mass) of local boids

With those 3 simple rules, a group of relatively disconnected individuals could form dynamic and useful group behavior. Reynolds’ simulation further proves that complex behavior emerges from simple conditions.

Boids could beautifully reproduce the complexity of flocking birds for three critical reasons:

  • Local rules – each boid had an innate ruleset that guided how it interacted with the boids around it
  • Simple rules – these rules were easy enough to parse in real-time, allowing the flock to quickly anticipate and adapt to change
  • Shared rules – these rules were shared across the population of boids, allowing them to quickly align their behavior

With that in mind, now consider how most managers try to guide collective action within organizations:

  • First, managers tend to start with top-down mandates instead of investigating or shaping the rules which actually guide individual behavior (mandates, historically, do very little to change individual behavior) – this results in disempowered teams or rogue actors
  • Second, managers tend to react to change by adding more and more rules (both implicit and explicit) until those rules become too unwieldy for individuals to follow effectively – this results in bureaucracy, slowed progress, and indecision
  • Third, managers (as organizations grow) tend to create silos that drift further apart, creating an absence of shared rules and cultures – this results in misalignment

Many management fads, when trying to guide organizational change, tend to fall into these same traps. Instead, we believe that you need to install simple and shared rules across all teams to manifest effective and responsive change.

We suggest starting with these 3 simple, shared, and local rules:

  1. Steer toward the customer: teams should serve either an external or internal customer and should reduce the distance between themselves and that customer (both in time and in physical distance)
  2. Steer toward autonomy: teams should be able to quickly concept and validate their own new ideas with little to no oversight or dependencies
  3. Steer toward alignment: teams should have cross-communication and cross-collaboration toward achieving their aligned objectives

Just as each boid constantly measured its progress and evaluated its ruleset, teams should routinely measure themselves against these rules (ie, Have we made progress toward greater autonomy this quarter?). Individuals should also make decisions with these simple rules in mind (ie, Will doing this steer us toward our customer or away?).

Of course, the only way to find out if these simple rules can guide your organization to faster and more effective change is to put them into practice.

Published January 13, 2019