Working definition
Social Sclerosis is the gradual hardening of an institution's norms, roles, incentives, relationships, and routines until adaptation becomes more threatening than underperformance.
Organizations often describe change as a problem of information or capability. If people understood the new strategy, received better training, or gained access to the right technology, the organization would move.
Sometimes that is true. Often it is not. People can understand the need for change and still preserve the system that prevents it. The obstacle is not ignorance. It is that the existing way of working distributes status, safety, authority, and responsibility. Changing the work also changes those arrangements.
A common scene
Consider a representative marketing team that introduces an AI workflow capable of producing a credible first campaign draft in two hours instead of two weeks. The demonstration succeeds. Leadership praises the efficiency. The work then enters the same sequence of reviews it followed before: channel lead, brand lead, product owner, legal, business-unit leader, and executive sponsor.
The production constraint has changed, but the decision system has not. Reviewers continue to protect separate risks, nobody owns the coherence of the final decision, and the organization measures the experiment by elapsed time. The conclusion becomes that the technology failed to deliver the promised speed.
This is a composite example, not a claim about one company. Its value is diagnostic: the new capability did not remove coordination costs, status boundaries, or fragmented authority. It made them easier to see.
AI makes the pattern easier to see
AI tools can make analysis, content, recommendations, and competent execution dramatically easier to produce. Yet many organizations respond by attaching the new capability to old approval chains, job boundaries, incentives, and assumptions about expertise.
The tool changes. The organization around the tool does not. Teams produce more without deciding better. Leaders ask for experimentation while punishing visible failure. Governance appears after adoption, ownership remains fragmented, and nobody knows whose judgment prevails when the machine and the experienced operator disagree.
This is why the hardest AI problems are frequently social systems problems. Technology can expose an institution's rigidity, but it cannot decide how authority, accountability, and trust should be redesigned.
The institutional evidence points in the same direction. The NIST AI Risk Management Framework treats governance as continuous and calls for explicit roles, responsibility, and human oversight. OECD workplace surveys similarly show that skills, cost, and trust shape AI adoption. Capability enters an existing social arrangement; it does not replace one.
Four ways a system hardens
- Roles become identities. A change in workflow is experienced as a change in personal worth or professional status.
- Routines become proof of competence. The organization continues a familiar process because abandoning it would call earlier decisions into question.
- Metrics preserve the old game. Leaders ask teams to adopt new capabilities while rewarding the outputs and behaviors of the previous system.
- Responsibility becomes diffuse. Everyone participates in the change, but nobody owns the decision or the consequence.
These mechanisms reinforce one another. A role becomes part of professional identity; the routine associated with that role becomes evidence of competence; the metric rewards preserving the routine; and shared participation makes responsibility difficult to locate. What looks like individual resistance can therefore be a rational response to the incentives and risks created by the system.
Marketing is a useful laboratory
Marketing sits at the intersection of technology, creativity, measurement, persuasion, customer behavior, and organizational politics. That makes it one of the first places Social Sclerosis becomes visible.
A team may gain the ability to generate fifty campaign concepts in an afternoon while retaining a decision process built for five. Content output increases while positioning becomes less coherent. New dashboards appear while the organization avoids choosing which measures matter. AI accelerates production, but the unresolved social system absorbs the gain.
The constraint was never merely the cost of production. It was the institution's ability to decide, align, and act.
Questions for leaders
- Which current routine exists because it still works—and which exists because changing it would threaten someone's status or certainty?
- Where have we introduced a new tool without changing decision rights?
- What behavior do our incentives reward, regardless of what our strategy says?
- Who owns the outcome when human and machine judgment conflict?
- What would we stop doing if our priorities were genuinely clear?
The practical implication
Social Sclerosis is not an argument against institutions or routines. Organizations need stability. The problem begins when stability stops serving the purpose and starts protecting itself.
The counterargument is important: resistance is not always irrational sclerosis. A slower team may understand customer risk, legal exposure, or operational dependencies that an enthusiastic change agent has missed. Speed is not proof of adaptation, and delay is not proof of dysfunction.
The diagnostic question is whether the objection can be made explicit and tested. Healthy caution names the risk, identifies the evidence required, assigns an owner, and states the conditions under which work may proceed. Sclerosis keeps the current arrangement beyond examination or moves the standard each time the evidence arrives.
The executive task is therefore not simply to introduce better technology. It is to make the surrounding system discussable: the assumptions, incentives, roles, relationships, and decisions that determine whether the technology can become useful.