Why systems literacy—not merely technical literacy—may be the essential skill of the AI age

Jodi Schiller

Judy Grupenhoff’s essay, “If You Think in Systems and See Patterns That Others Don’t, This Is Your Moment”, offers a powerful recognition of minds long dismissed as scattered, nonlinear, or difficult to follow. She argues that people who naturally hold multiple streams of information at once—and perceive the relationships among them—may be unusually well suited to this moment of technological and social complexity.

When I read her essay, I recognized the path she described. I began in theater, moved into psychology, then augmented and virtual reality, and now artificial intelligence. From the outside, these may look like separate fields. To me, they have always been connected: each explores how human reality is constructed, perceived, and shaped. That same systems orientation informs my analysis of unpaid care labor, economic dispossession, domestic violence, homelessness, legal inequality, and institutional abandonment—not as isolated problems, but as interacting parts of systems that transfer their costs onto the people with the least power.

Grupenhoff identifies the people who naturally think this way. But her insight points toward a larger transition. Systems thinking can no longer remain an unusual cognitive gift possessed by a few people who instinctively see the whole. In the age of artificial intelligence, it must become a shared literacy.

From Systems Thinkers to Systems Literacy

For centuries, institutions have trained us to divide problems into smaller parts, assign each part to a separate discipline, and optimize each component independently. That approach has produced extraordinary advances. But it also creates a dangerous illusion: that a system will improve whenever one measurable part of it improves.

Systems do not behave that way.

A system is not merely a collection of components. It is the relationships among them: the feedback loops, incentives, delays, dependencies, adaptations, and transfers of cost that produce outcomes no individual part was designed to create. Change one element and other elements respond. Solve a problem in one location and the burden may simply reappear somewhere else. Improve the metric and people may adapt their behavior to the metric rather than to the purpose it was intended to serve.

This is why first-order thinking is no longer enough. First-order thinking asks: Will this action produce the immediate result we want? Systems thinking asks: And then what?

Who will change their behavior in response? What feedback loop will be created? What cost will be displaced? Who will receive the benefit, and who will absorb the externality? What may happen months or years later, after the intervention has traveled through institutions and lives?

These are not abstract academic questions. They are the practical questions that determine whether an apparent solution actually solves anything.

AI Is Not Outside the System

Artificial intelligence makes systems literacy urgent because AI is often presented as a tool applied to a problem from the outside. But AI never remains outside the system it enters. It changes the system.

Introduce AI into a workplace and it changes more than productivity. It may change hiring, training, supervision, expectations, bargaining power, professional judgment, and the distribution of responsibility when something goes wrong. Introduce it into education and it changes more than how students complete assignments. It may change what it means to learn, what teachers evaluate, how knowledge is demonstrated, and who has access to personalized intellectual support. Introduce it into law, medicine, public benefits, or policing and it may alter not only decisions, but also who is seen, believed, classified, monitored, or excluded.

The tool becomes an actor in the environment. Human beings adapt to it. Institutions reorganize around it. Incentives change. New dependencies develop. An intervention introduced to improve one task can transform the entire structure in which that task occurs.

That is why it is not enough to ask whether an AI system produces an accurate answer. We must ask what happens when people begin trusting that answer, when institutions build procedures around it, when workers are evaluated through it, and when those affected by it have no meaningful way to challenge it.

AI Does Not Relieve Us of Systems Thinking

AI is remarkably capable of finding relationships across large bodies of information. It can help us hold multiple streams of thought, detect recurring structures, test possibilities, and translate nonlinear thinking into language that others can follow. This is the liberating possibility Grupenhoff describes: AI as a partner that helps systems thinkers communicate without flattening their thinking first.

But AI does not automatically understand a system simply because it can describe one. Pattern recognition is not the same as causal understanding. A fluent answer is not necessarily sound judgment. AI can reproduce the assumptions embedded in its information, optimize the objective it is given while neglecting what was left out, and offer a convincing recommendation without bearing any responsibility for its consequences.

The more powerful the technology becomes, the more important human systems judgment becomes—not less.

To work responsibly with AI, we must learn to specify not only the task, but also its context. We must identify the actors, incentives, constraints, histories, power relationships, and possible consequences surrounding the task. We must examine what the proposed solution counts, what it excludes, and what it treats as someone else’s problem.

In other words, working well with AI requires us to think beyond the prompt. We must think about the system into which the answer will be released.

What Broken Systems Hide

Many institutional failures persist because each institution sees only its own fragment.

Unpaid care labor is treated as separate from the market economy, even though the economy depends upon it. Domestic violence is treated as a private incident, while its consequences reappear in housing, employment, health, education, and the courts. Homelessness is addressed as a condition visible on the street rather than as the downstream result of failures in family systems, wages, housing, health care, law, and public administration. Legal proceedings may be judged procedurally correct even when profound inequalities in money, information, representation, and safety determine the outcome.

Each institution can claim to have handled its assigned piece while the person moving through all of those institutions experiences one continuous system.

This fragmentation also conceals responsibility. When harms are divided into separate administrative categories, no institution has to recognize the whole trajectory. Costs created in one system are absorbed by another—or, most often, by the individual with the least power to refuse them.

Systems thinking makes those transfers visible. It asks us to follow the consequence rather than the organizational chart.

The Literacy This Moment Requires

Systems literacy does not mean that everyone must become a systems theorist. It means learning a basic discipline of attention:

• Look for relationships, not only objects.

• Trace feedback loops, not only immediate causes.

• Expect delays between an intervention and its full effects.

• Distinguish a shifted burden from a solved problem.

• Ask what the measurement excludes.

• Examine how power determines who receives benefits and who absorbs costs.

• Consider second- and third-order consequences before acting at scale.

• Revisit the system after people have adapted to the change.

These habits are essential when working with AI because AI dramatically increases our capacity to intervene. It allows individuals and institutions to generate, classify, recommend, automate, and scale at unprecedented speed. But faster intervention without deeper systems awareness can accelerate failure as easily as progress.

We are therefore entering a paradoxical moment. AI can help make systems visible, but it can also deepen systemic harm. It can connect knowledge across disciplines, but it can also automate the blindness already built into institutions. It can help previously misunderstood thinkers communicate, but it can also amplify the authority of systems that refuse to listen.

The outcome will depend partly on whether we learn to see beyond the immediate output.

Grupenhoff is right that this may be the moment for people who have always thought in systems. But it must also become the moment when systems thinking moves from the margins into ordinary public understanding. The future will not be shaped only by those who know how to operate AI. It will be shaped by those who understand what happens after AI is operated—when its outputs enter human institutions, encounter unequal power, trigger adaptation, and produce consequences far beyond the original prompt.

The essential question of the AI age is not simply, What can this technology do?

It is: What system are we changing—and what will that system do in response?