IB Digital Society 2.6A Systems and System Thinking Notes

2.6A Systems and System Thinking

Systems as a Concept

1. Systems is a concept with a way of looking called systems thinking. A system is a set of parts that interact and depend on one another to serve a purpose; study the whole and its connections, since the parts alone will not explain it.
2. Meadows: any system is three things, its elements, the interconnections between them, and a purpose. Elements are easiest to see; interconnections and purpose explain how it behaves.
3. A social platform as a system: elements = users, posts, ranking algorithm; interconnections = rules deciding who sees what; purpose (owner-set) = maximise engagement. Change the purpose and everything shifts at once, which is why it is the most powerful yet least visible part.
4. Systems thinking asks how the parts affect each other and what the whole produces. Almost everything in digital society, from a platform to a supply chain, is best read this way.

Feedback Loops

1. Systems are steered by feedback loops, where the output loops back to change the input. Two kinds.
2. Reinforcing loop amplifies change (more leads to more): a shared post is shown to more people who share again, snowballing into a viral trend. Balancing loop pushes back to a target, like a thermostat: moderation or a rate limit corrects a platform when it drifts too far.
3. A reinforcing loop in a recommendation feed narrows and deepens each turn (one video, then three, then ten). Unchecked it runs to an extreme unless a balancing loop, or a human, reins it in.
4. Two things make loops hard to manage: delay (an effect arrives long after its cause, so people over-correct) and tipping points (a balancing loop gives way and a reinforcing one takes over, flipping the system into a new state).

Emergence: The Whole Is More Than the Parts

1. Emergence: many parts interacting produce behaviour no single part has. A bird flock, a viral post, a market price, an online pile-on, are patterns from many small interactions, designed by no one participant.
2. Digital example: no engineer programmed a market to collapse, yet in the 2010 flash crash automated traders reacting to one another sent share prices plunging and recovering in minutes. The behaviour belonged to the interacting whole.
3. Emergence is why you cannot fully predict a system from one user or one line of code; outcomes can surprise even the builders.

Intended and Unintended Consequences

1. The signature move of the concept: separate what a change was meant to do from what it actually did. Interconnected parts mean a change ripples outward into effects no one planned.
2. A classic pattern is the perverse incentive, where rewarding one measure encourages the wrong behaviour (the cobra effect).
3. Example: a platform boosts "meaningful interactions" to get friendlier conversation but rewards outrage, because angry posts draw the most replies. The same change produced both the intended aim and the harm.
4. Lesson: watch for the ripple. A fix will have effects on the whole, so name the likely side effects before making the change.

Interdependence and Cascading Failure

1. Interconnection makes systems powerful. It also makes them fragile: a failure in one place can cascade through the whole, a risk that grows as systems become interdependent.
2. Digital systems tie into physical ones, so a digital failure has physical effects. Shared cloud providers, payment networks, and login systems create single points of failure that can take down many things at once.
3. Outage example: when a single cloud region or major content-delivery network fails, thousands of unrelated sites, apps, and smart-home devices go dark at once, because they all quietly depended on it.
4. Systems thinking asks you to see the links clearly. The links that carry a cascading failure also carry help and repairs. Resilience (slack, backups) is as much a values and ethics and power question as a technical one.