Good question Dan.
A lot of methods and algorithms used in AI/ML are inspired by analogues observed in nature.
Example: Collective Animal Behavior (CAB)
… think a school of fish
… a swarm of bees
… flocks of birds
By observing these patterns in nature and understanding what causes the behaviors in response to environmental conditions or outside inputs we can apply that understanding to events happening online.
We then compare and consider if the same responses would elicit different results based on smaller groups or individuals, and are appropriate for the perceived threat?
Example: A school of fish reacting to predator, say a shark, and initiate a startle cascade — versus how they react to bubble net feeding being coordinated by whales.