Homeostats for a Risk Society: Relational Systems and Psychological Safety in Practice
I was invited to write a piece for Risk Society, a knowledge co-production platform and expert network that bridges academic research, policy development, and real-world practice through interdisciplinary dialogue on public risk management and anticipatory governance. They bring together candid insights across diverse fields, to advance innovative strategies for navigating complexity, building societal resilience, and strengthening anticipatory capabilities in an uncertain world.
In contemporary risk society, the dominant failure mode is no longer a lack of frameworks or models; it is the inability of human systems to surface weak signals, act on uncertainty and adapt at the speed of emerging threats. For an audience steeped in risk management, relational public services and psychological safety are best understood as core features of the control environment and feedback architecture of resilient social systems, not as “soft” add‑ons.
Relational public services reconfigure the public sector from a chain of discrete transactions into a network of enduring relationships with citizens, communities and frontline professionals. Operationally, this shifts emphasis from throughput and narrow KPIs toward context, continuity and local intelligence. For risk managers, the payoff is obvious: richer, earlier and more granular data about emerging exposures. A practitioner who knows a family, community or a workplace over time acts as a sensemaker in context, able to detect pattern breaks – financial stress, coercive behaviour, climate‑related vulnerability, deteriorating mental health – long before they trigger formal incidents or claims. Relationally based supports therefore increase the sensitivity of the system’s “detection layer” and widen its repertoire of mitigations through community assets, informal networks and cross‑agency collaboration. As a conduit for communicating risk, it relies on detecting signals as much as having a safe route through which to process them onwards.
Psychological safety is the cultural condition that determines whether that detection layer actually transmits usable signals. It is far more than the ability to take risks or innovate as the current transactional framing in the literature suggests. In reality, psychological safety is a systemic construct: it is about feeling safe enough to engage with complexity, rather than perform for the sake of the stats. It is about turning up for people, and being present enough to stay with uncertainty, without trying to fix them from a deficit-based analysis of need. Safety is about connection: connection is about mattering, mattering leads to belonging. This is the nuts and bolts of building trust: it also determines whether that detection layer actually transmits usable signals. In risk terms, it is a property of the control environment that governs error reporting, escalation behaviour and policy-maker’s real‑world appetite for bad news. Where psychological safety is low and trust is absent, near misses, weak signals and dissenting assessments are systematically filtered out: staff self‑censor, communities disengage indicator lag behind. The result is latent risk accumulation and surprise failures. Where psychological safety is high, people at every level are more willing to raise concerns, challenge assumptions and admit uncertainty. That dramatically improves the volume and quality of upward and lateral risk information, and it enables timely adjustment of controls and strategy.
Taken together, relational public services and psychological safety function like a homeostat in complex, high‑uncertainty environments. They create dense sensing networks (relationships) and low‑friction feedback channels (psychological safety) that allow the system to maintain wellbeing, social cohesion and continuity within acceptable bounds despite external shocks. For risk management this is crucial: in a world of systemic, non‑linear threats such as climate change, pandemics, cyber‑risk and social unrest, ex‑ante control design and static risk registers are necessary but insufficient. What differentiates resilient systems is their ability to learn in real time, reconfigure resources and even revise objectives as conditions change.
In practice, this suggests three implications for risk professionals:
Treat relational capability as a core component of design, implementation and evaluation of social programmes.
Explicitly name a systemic rather than instrumental understanding of psychological safety as a risk factor.
Stress tests should examine not just infrastructure and procedures, but also whether the system’s human networks and culture can sustain honest, timely feedback under stress.
By integrating these dimensions into resilience planning, we rehumanise our systems: relationships become a critical resource that determines the adaptive capabilities that insure us against future shocks.
