Why linear policy planning fails on wicked problems like youth social media regulation — and what to do instead.
Governing a bridge and governing a social media platform are different kinds of problems, even though both look like “engineering” from a distance. A bridge is complicated: expert analysis produces a stable, predictable solution. A platform is complex: the moment you regulate it, users, algorithms, and lawmakers all adapt around the new rule, changing the system you were trying to fix. This article combines five frameworks — the Cynefin model, systems thinking, bounded rationality, nudge theory, and Actor-Network Theory — into one adaptive approach for governing problems of the second kind, and applies it to a real 2026 court case over youth social media law.
Complicated vs. Complex: Why the Old Playbook Breaks
For decades, governance relied on a simple assumption: analyze the problem, have experts design a solution, implement it, and expect predictable results. That works for complicated problems — building a bridge, auditing financial records — where cause and effect are stable enough for expert analysis to clarify a defensible answer.
It breaks down on complex problems, where variables interact recursively and the very act of intervening changes the system. Regulating platform safety, adolescent mental health, or generative AI all fall into this category.
| Complicated Domain | Complex Domain | |
|---|---|---|
| Sequence | Problem → Expert Analysis → Clear Solution | Problem → Probe/Experiment → Sense/Feedback → Adapt |
| Appropriate response | Sense–Analyze–Respond | Probe–Sense–Respond |
Public administration scholars Jos Raadschelders and Travis Whetsell sharpen this diagnosis: complex public problems are as much epistemic as technical — resolving them means managing competing claims about what counts as knowledge, who holds authority, and which experiences matter, not just running better analysis.
The Four Knowledge Sources
Raadschelders and Whetsell identify four distinct knowledge streams that any complex digital policy decision has to reconcile:
| Knowledge Source | What It Is | Digital Policy Example |
|---|---|---|
| Scientific / Disciplinary | Peer-reviewed research, formal standards, statistical evidence | Empirical studies on teen screen time and circadian rhythm disruption |
| Organizational | Institutional memory, corporate routines, operational strategy | Platform engagement algorithms and retention targets |
| Societal | Cultural norms, public values, social movements, ethics | Parental concerns over digital autonomy and youth wellbeing |
| Individual | Lived experience, personal habit, daily practical judgment | An adolescent’s daily interactions with notification prompts |
When these four sources clash adversarially rather than collaboratively, regulatory stalemates follow. Synthesis, not suppression of one source by another, is what moves a problem forward.
Five Frameworks for Socio-Technical Governance
The Cynefin Framework. Developed by Dave Snowden, Cynefin helps leaders diagnose which kind of problem they’re actually facing. In complex domains, the move is from Sense-Analyze-Respond to Probe-Sense-Respond: initiate small, safe-to-fail experiments, observe the reaction, and adapt.
Systems Thinking & Feedback Loops. Socio-technical systems run on feedback loops. A reinforcing loop accelerates platform engagement (more data → better personalization → higher salience → longer use); interventions need to find leverage points that create balancing loops without triggering unintended displacement elsewhere.
Bounded Rationality & Satisficing. Herbert Simon’s foundational insight: decision-makers operate under real limits of time, attention, and cognitive bandwidth. Waiting for perfect data in a fast-moving system is often worse than satisficing— selecting a defensible, adequate path that generates immediate learning.
Choice Architecture & Nudge Theory. Platforms aren’t neutral conduits; they’re carefully architected environments. Default settings, infinite scroll, and algorithmic recommendations all function as behavioral nudges. Regulatory counter-nudges — default break reminders, privacy-first settings — can reshape behavior while still respecting user autonomy.
Actor-Network Theory (ANT). Bruno Latour’s framework insists that non-human entities carry real agency in a system. Algorithms, smartphones, terms-of-service contracts, and data servers actively mediate human relationships and power dynamics — a digital platform is not merely a passive forum for speech, but an active mediator that translates incentives, attention, and identity into socio-technical arrangements.
Case Study: Youth Social Media and NetChoice, LLC v. Yost
Ohio’s Social Media Parental Notification Act (Ohio Rev. Code § 1349.09) requires platforms to secure verifiable parental consent before minors under 16 can create accounts.
[Image: network diagram showing State Law, Platforms/UX, and Minors & Parents as interlinked nodes, feeding up to Judicial Review (Sixth Circuit).]
The state’s intent: interrupt the reinforcing attention loop to protect adolescent mental health — citing Pew Research’s 2025 finding that 45% of U.S. teens say they spend too much time on social media, up from 36% in 2022.
The industry’s response: NetChoice challenged the law, arguing it chilled minors’ protected First Amendment expression.
The judicial outcome: in NetChoice, LLC v. Yost, the Sixth Circuit reversed a district court injunction, finding NetChoice had failed to establish the law was facially unconstitutional.
The governance insight: treating platform access as a static legal binary — speech versus ban — misses the underlying socio-technical architecture. Durable digital policy has to address the behavioral nudges, data-extraction loops, and interface defaults simultaneously, not just the legal question in isolation.
A Connection Worth Naming
Cynefin’s Probe-Sense-Respond sequence for institutions is structurally the same move this site’s own habit loop asks of an individual: don’t assume you already know the answer — gather signal first, then act. Pause → Verify → Contextualize → Participate is Probe-Sense-Respond scaled down to one person evaluating one claim; Cynefin is that same discipline scaled up to an institution evaluating a policy. Both reject the same failure mode: treating a fast-changing, interactive system as if it were static and fully knowable.
Adaptive Governance, Applied Beyond Platform Law
This framework extends past public regulation into enterprise AI governance, where deploying a large-scale model is rarely a purely technical task — it’s an organizational and socio-technical deployment shaped by institutional workflows, legal obligations, and situated human expertise just as much as by model performance.
| Step | Action |
|---|---|
| 1. Diagnose the Domain | Use Cynefin sense-making to separate complicated technical fixes from complex socio-technical risk. |
| 2. Map the Epistemic Landscape | Integrate scientific metrics, organizational goals, legal/societal norms, and user feedback. |
| 3. Deploy Safe-to-Fail Probes | Launch narrow, reversible, empirically monitored initiatives — privacy-protective defaults, pilot AI safety guardrails, time-limited feature rollouts. |
| 4. Sense Systemic Feedback | Continuously audit for feedback loops, unintended behavioral shifts, compliance costs, and displacement. |
| 5. Orchestrate Adaptive Responses | Fine-tune code, institutional workflows, and policy guidelines in tandem, rather than treating any one as fixed. |
The same five-step logic applies to a food-price crisis as it does to platform law or AI deployment — diagnose the actual constraint before choosing the instrument, and build in a way to sense whether the instrument is working.
Conclusion: Agility as Institutional Rigor
Rigid predictability is an illusion in a system defined by rapid technological and social change. Effective digital policy — legislative or corporate — depends on institutional humility, continuous learning, and the willingness to synthesize multiple knowledge sources rather than defaulting to whichever one is loudest. When confronting a genuinely wicked socio-technical problem, agility isn’t the absence of rigor; it’s the specific form rigor has to take.
Download
References
- Bhargava, V. R., & Velasquez, M. (2021). Ethics of the attention economy: The problem of social media addiction. Business Ethics Quarterly.
- Faverio, M., Anderson, M., & Park, E. (2025). Teens, social media and mental health. Pew Research Center.
- Latour, B. (2005). Reassembling the Social: An Introduction to Actor-Network-Theory. Oxford University Press.
- National Institute of Standards and Technology. (2023). Artificial Intelligence Risk Management Framework (AI RMF 1.0). U.S. Department of Commerce.
- Raadschelders, J. C. N., & Whetsell, T. A. (2018). Conceptualizing the landscape of decision making for complex problem solving. International Journal of Public Administration, 41(14), 1132–1144.
- Simon, H. A. (1957). Models of Man: Social and Rational. John Wiley & Sons.
- Snowden, D. J., & Boone, M. E. (2007). A leader’s framework for decision making. Harvard Business Review, 85(11), 68–76.
- Thaler, R. H., & Sunstein, C. R. (2008). Nudge: Improving Decisions About Health, Wealth, and Happiness. Yale University Press.
- NetChoice, LLC v. Yost, No. 25-3371 (6th Cir. June 18, 2026).
- Brown v. Entertainment Merchants Association, 564 U.S. 786 (2011).