Technology keeps reorganizing civilization faster than civilization can respond. This is not a new problem. It is a systems problem. And it finally has a science.
In 1963, a group of mid-century intellectuals gathered at the University of Michigan to ask a question that turned out to be premature by sixty years. The volume they produced, World Technology and Human Destiny, included contributions from the political philosopher Raymond Aron, the physicist Robert Oppenheimer, and the diplomat George Kennan. It is a careful, serious book, and it has aged badly in exactly the way that correct diagnoses age badly: by becoming more obviously right with every passing year while the problem it identified remains unsolved.
The question they were asking was structural, not moral. They were not asking whether technology was good or bad. They were asking what technology does to the social order — how it reorganizes institutions, values, power, and human behavior in ways that outpace the capacity of the institutions, values, and norms designed to manage it. They had atomic weapons in mind, and the industrial reorganization of society, and the Cold War political systems that were already beginning to show the strain of managing unprecedented destructive capability with pre-modern political institutions.
Aron was the most precise. His central argument was that industrial society produces not just new tools but a reorganized social order — that you cannot understand what a technology does by examining the technology alone, because what matters is how it reshapes the whole system of interactions: work, power, ideology, and governance together. A new technology doesn't add a capability to an existing civilization. It changes the civilization.
That is a systems insight, stated in the vocabulary of political sociology rather than complexity science. But the underlying claim is identical to what systems scientists were formalizing in the same decade: that you cannot understand the behavior of a whole by examining its parts, that the interaction structure matters as much as the components, and that emergent behaviors at the level of the whole can be wildly different from anything the components would predict.
Aron and Oppenheimer were not systems scientists. They did not use the vocabulary of feedback loops, attractor basins, or phase transitions. But they were thinking in systems, and what they saw was real: that the most dangerous feature of powerful technology is not its direct effects but its capacity to reorganize the attractors of human civilization — the stable patterns of behavior, institution, and incentive that societies settle into and maintain — faster than the governance structures designed to keep those attractors healthy can track and respond.
They were right in 1963. The diagnosis has been correct for sixty years. The response has still not arrived at the necessary scale.
The reason is not ignorance. Plenty of people understand, in general terms, that social media is reorganizing political behavior, that AI is changing the economics of information, that automation is restructuring labor markets, that these changes are happening faster than political institutions can adapt. The problem is not awareness. The problem is that awareness of a systems-level problem, without the tools to analyze it as a systems-level problem, produces nothing but increasingly alarmed commentary.
This is where the scientific lineage that Aron and Oppenheimer were working beside — but never quite with — becomes essential.
While Aron was writing political philosophy and Oppenheimer was warning about civilizational unpreparedness, a parallel scientific tradition was developing the formal tools for understanding exactly what they were describing. Norbert Wiener's cybernetics had already established, in 1948, that feedback — the process by which a system's outputs feed back to shape its own behavior — was the organizing principle of purposive behavior in both machines and organisms, and that a system with broken feedback would behave pathologically regardless of how capable its individual components were. Jay Forrester's systems dynamics, developed through the 1950s and 1960s, showed by simulation what Aron was arguing by analysis: that systems with the same feedback structure produce the same behaviors regardless of what their components are made of, and that the most intuitively appealing interventions are often the ones that make the underlying dynamics worse.
The Santa Fe tradition, beginning in the 1980s, formalized complex adaptive systems — systems composed of agents that learn and evolve, producing emergent patterns at the aggregate level that no individual agent intended or could predict. Network science, at the turn of the millennium, added the architectural dimension: how components are connected shapes behavior as profoundly as what those components do. Resilience science showed how systems absorb disruption — or fail to — and how they sometimes shift suddenly and irreversibly into qualitatively different configurations.
By the early twenty-first century, the science of complex systems had produced a rigorous account of exactly the phenomenon Aron and Oppenheimer were describing. It had a name for what technology does to civilization: it shifts attractors. It had a framework for why institutions lag: they are themselves attractors, maintained by feedback loops that resist change. It had a vocabulary for the dangerous moments: threshold dynamics, tipping points, phase transitions — the points at which a system that appeared stable suddenly reorganizes into something new, sometimes without any possibility of return.
The two traditions — formal systems science and humanist civilizational criticism — had been converging on the same insight from opposite directions for sixty years without ever fully combining. The gap between them is not a footnote in intellectual history. It is a structural feature of the problem itself: the very disciplinary fragmentation that systems science identifies as a characteristic failure mode of complex institutional systems has kept the diagnosis and the tools for responding to it from being integrated at the scale the problem requires.
Now consider the current moment.
Social media is not a neutral communication channel. It is a feedback machine that selects for and amplifies what spreads — which turns out to be, with reliable consistency, outrage, tribal signaling, status anxiety, and compulsive self-display. This is not because social media companies intended to degrade public discourse. It is because engagement metrics create a reinforcing loop: content that triggers strong emotional response gets amplified, amplification produces more of that content, the content shapes what users come to expect, and the expectations drive the design choices that produce more of the same content. No one is steering. The system is running its feedback dynamics, and the attractor it is stabilizing — fragmented, polarized, manipulable — is genuinely dangerous to every governance institution that depends on a minimally coherent public sphere.
This is the same pattern Aron described for industrial capitalism's effect on political ideology. New capability; reorganized attractor; institutional lag; nobody clearly in charge of the overall dynamics.
Artificial intelligence is the same pattern operating on cognition itself. The cost of generating plausible text, persuasive argument, and synthetic media is dropping toward zero. This means that the information environment — the medium in which the epistemic institutions of democracy (journalism, peer review, public deliberation) operate — is being flooded with generated content that those institutions were not designed to navigate. The feedback loops that allowed truth to surface slowly above noise were calibrated for a world in which generating high-quality fake content was expensive. That world is gone. The institutions have not yet adapted. The attractor they were designed to maintain — a minimally shared epistemic commons — is under pressure from a force whose scale was unthinkable five years ago.
Robotics completes the pattern by projecting it into the physical world. What has so far been a reorganization of information, attention, and institutional incentives becomes a reorganization of material reality: supply chains, infrastructure, security, and the physical environments in which human life is conducted. The same feedback instabilities that produce epistemic fragility in an AI-saturated information environment can produce physical fragility in an automated material world, with failure modes that are harder to observe before they cascade and harder to reverse once they do.
The key feature in each case is not the technology itself but the structural relationship: expanding capability, lagging stewardship, and a system of incentives that rewards the expansion without requiring the stewardship. Each technology is an attractor-shaping force — not a tool that humans choose whether to use, but a system that reorganizes the environment in which human choices are made. It changes what is easy, what is rewarded, what is expected, and what is visible. The civilization that has access to social media, AI, and autonomous systems is not the same civilization navigating the same problems with better tools. It is a different civilization, with different attractors, some of which are deeply unhealthy.
Aron's warning, restated for 2026: you cannot understand what is happening by examining any of these technologies alone. What matters is how they interact with and reorganize the whole system of institutions, incentives, values, and behaviors. And that whole-system behavior will not be visible to any analysis that examines the parts separately.
The right response to this is not nostalgia, prohibition, or technological optimism. It is not more regulation drafted by people who do not understand feedback dynamics. It is not more alarmed commentary by people who understand the problem but lack the tools to specify an intervention.
It is systems design.
Systems design means starting with the dynamics, not the surface behavior. It means asking what attractors the current technological configuration is stabilizing, what feedback loops are maintaining those attractors, how deep the basins are, and where the leverage points lie — the places where a precisely calibrated intervention can shift a stable pattern without provoking the system's full self-stabilizing response. It means designing governance structures that have requisite variety — enough internal complexity to match the complexity of the systems they are governing. It means building feedback into every intervention so that it can detect and correct its own trajectory as the system responds.
None of this is exotic. These are principles that ecologists, engineers, and epidemiologists use routinely. They are not being applied, at the necessary scale and with the necessary rigor, to the governance of technological civilization. The reason is institutional: the disciplinary structures that organize knowledge production are themselves a pre-complexity artifact, dividing the study of exactly the phenomena that can only be understood whole. Political scientists study political institutions. Economists study economic behavior. Sociologists study social dynamics. Computer scientists study AI. None of these disciplines, working alone, can see the attractor landscape that all of them are operating in.
This is the gap that systems science exists to close. Not by replacing the disciplines, but by providing the integrative framework — the vocabulary, the modeling tools, and the intervention methodology — that allows disciplinary knowledge to be brought to bear on the whole-system dynamics that none of the disciplines can see alone.
This is the gap ISRD works in.
The methodology applied here is attractor engineering: the systematic process of identifying what holds a harmful system pattern in place, finding the leverage points where it is most susceptible to movement, and designing intervention sequences that shift it toward a healthier configuration.
The problems worth working on — chronic disease that persists because the physiological systems maintaining it are organized to maintain themselves; social and political fragility that persists because the institutional structures that could resolve it are themselves stuck in dysfunctional attractor states; technological governance that fails because the governance systems are calibrated for a complexity level the technologies have long since exceeded — all have the same structural feature. They are not problems that more information, more resources, or more compliance will solve. They are attractor problems. And attractor problems require attractor engineering.
The science between the sciences is finally mature enough to apply. The capability-stewardship gap that Aron and Oppenheimer saw in 1963 is wider now, and the systems producing it are moving faster. But the tools for seeing it clearly, diagnosing it precisely, and designing interventions that actually shift it — those tools exist.
Deploying them is the work.