The Forever Crisis: Adaptive Global Governance

Adam Day. 2025. The Forever Crisis: Adaptive Global Governance for an Era of Accelerating Complexity. London and New York: Routledge. 184 p.

Adam Day, as Head of the UN University Centre for Policy Research in Geneva, who co-led the Secretariat of the UN High-Level Advisory Board on Effective Multilateralism, could not be better placed to give us a new constructive perspective on the challenges of global governance. The following is a summary of the main points of his book without the extensive examples and academic references. As a complex systems scientist myself, this work deeply resonates with me.

Our traditional patterns of thinking, often linear from cause to effect, and framed within one of many silos in a world of increasing specialisation, have failed to help us turn the corner in the face of many complex challenges, from the threat of nuclear war to climate change and artificial intelligence (AI). Governments agree to beautiful plans like the Sustainable Development Goals, but fail on implementation. The science of complex systems offers a constructive alternative, as this book demonstrates.

It provides a simple guide to complexity, applies this to global governance, and illustrates this with the approach to environmental challenges, violence and war, cyberattacks, AI, and saving the world in the future.

Systems science describes the characteristics of a complex adaptive system. Such a system involves many actors or agents that self-organise without any central leader, and are open to other systems or outside conditions. They do not follow fixed rules, but respond to probabilities, demonstrating emergent properties that are more than and different to the sum of individual actions. Relationships between local agents are the foundation of self-organisation, passing information through feedback loops in non-linear fashion, perhaps passing some threshold or tipping point that can be difficult to anticipate. Systems generally operate in cycles, and work across scales of space and time. They may maintain some attractive or steady state until pushed beyond some limit, and may experience recurrent volatility or even self-organised criticality and collapse. Change happens over time, but not smoothly, following punctuated equilibria of stability followed by rapid change. Change in complex systems is the result of the interaction between the form of the system and specific events or actions. Such systems evolve patterns and contours that shape our actions within them, and require effort to overcome. We cannot really control complex adaptive systems, especially top-down, but we can try to exert influence or affect the underlying pattern, since relationships drive change. Change has to fit across different scales with multilevel adaptive governance. Such systems resist change with resilience, which can be positive to maintain stability, or negative when change is desirable, but it is too easy to give up when change seems insurmountable.

Day then applies complexity to global governance, since governance processes to manage global issues also constitute complex systems and should be addressed with the tools of complexity thinking. In a forward-looking application of complexity thinking, he offers viable options to those directly involved in shaping global governance. Across all the definitions of global governance, he extracts its key characteristics. It is governance, not government, since no single global authority can decide for everyone, but rules and constraints on behaviour should shape how actors relate to each other, including collective actions to manage common resources. Governance has both an institutional and a functional aspect, and from a systems perspective is better considered a verb rather than a noun. With globalisation and our increasing digital existence, the system is complex with many overlapping networks of actors and processes. While states are central to many global institutions, they are really nodes in networks of many different players, just parts of a system. While relations among states are defined as international, global refers to managing our complex and interdependent planetary system. Global governance is clearly relevant for the environment, finance, health, security, and technology, even when linked to other issues, and these are largely undergoverned spaces. They have shortcomings in enforcement, attribution and regulation. Day summarises today’s polycrisis as a rapid growth in undergoverned spaces.

The response is to move towards adaptive global governance, looking for effective responses in uncertain conditions. Environmental governance has pioneered this approach, with formal and informal institutions, experimentation and constant learning. The four core elements of adaptive governance are (1) mapping the dynamics and underlying patterns of complex systems; (2) designing organisations that can adapt to shifting circumstances; (3) encouraging multiple stakeholders in decision-making; and (4) feeding the experience back into future iterations of governance. This could be modelled as an Act-Sense-Decide-Adapt cycle. Adaptation can lead to the recognition that the system itself must be transformed, in a deeper systemic change, a more radical shift in the underlying rules, patterns and attractors. There are challenges because of the uncertainty about objectives, context and effectiveness when the imagined future is unknown, the stakeholders need to be determined, and confirmation biases avoided.

Day proposes a framework for adaptive global governance design, with flexible steps and questions to help in understanding how a given system works, what objectives make sense, and whether a proposed course of action is likely to achieve objectives. The first step is to map the system, including how actors relate to each other, what the key characteristics of the system are including its density and interconnectedness, how change happens within the system, and how all this defines the range of potential outcomes at a given moment. The second step is to orient towards goals, while adapting goals continuously. Setting meaningful goals, which could be environmental resilience, goal-setting as in the Sustainable Development Goals, or rule-making, while being demand-driven and flexible. The third step is to design a “fit” response as in the Darwinian definition of fitness through natural selection, testing and discarding ineffective approaches. Ways to determine the match between the problem and the solution can be spatial for geographic fit, temporal for the timespan of change, or flexible in addressing unforeseen changes. Complexity thinking suggests that governance responses can be less about fixed institutions and more about activation, linking and orchestrating across levels. The fourth step is to test robustness and resilience together, with a resilient system absorbing shocks and a robust system delivering in conditions of uncertainty, both forward and backward in time. The fifth step is to repeat through feedback loops, passing information back into the system and adapting based on the changed circumstances. The feedback can suggest the need for more fundamental transformation in the system, while considering the potential impacts of such a radical transformation. The evolution of adaptive governance, like evolution in nature, goes through experimentation, selection/adoption, and inheritance, and we can shape all three stages if we overcome habit and ideology, and encourage the diversity that allows adaptation.

The first case study in the book is environmental governance in the Anthropocene, where we are beyond the tipping point. Now that human activity is the dominant cause of environmental change, we have crossed six of nine planetary boundaries, causing the planet to cross irreversible thresholds including climate breakdown and the sixth extinction. Our linear thinking expects change to be gradual, but socioecological systems are discontinuous, abrupt and surprising. We cannot reverse a complex system, only anticipate changes, evolve alongside them, and shape the transition in real time, and we are already in a transformational moment. We need a system of governance that transforms us alongside our planet.

The chapter reviews the extensive academic work on environmental governance in a complex systems perspective, as this reflects the scientific understanding of ecological systems, but this is largely divorced from policy-making and has little impact despite the urgency. It offers complexity thinking as a bridge between different policy expertise areas and the decisions policymakers must make today, and proposes some of the more ambitious initiatives that have not yet been considered seriously. They go beyond trying to fix a dysfunctional machine, assuming that greater understanding will lead to action, creating a single global environmental governance, or trying to reverse engineer climate change and biodiversity loss. They may also be more immediately acceptable than making ecocide a crime or replacing capitalism.

The first idea is a world body with teeth, beyond voluntary commitments. Such a Global Environmental Organization could consolidate existing environmental obligations into a coherent global legal framework enforced by a global judicial body, able to handle environmental disputes, track progress, and hold actors accountable. The second is an “IPCC for the planet”, a science-policy interface to understand our impacts on the planet, how those impacts interact, and what policymakers can do to address them holistically. The third is to fix the global economy, orienting the international financial architecture to address the planetary crisis, exponentially increasing the scale of investment. The fourth is to “stuff Pandora back into the box” with technological interventions in the Earth’s natural systems, such as carbon dioxide removal, solar radiation modification or highly controversial geoengineering.

These proposals might make a difference, but they are problematic because they only try to fix part of the problem rather than addressing planetary socioecological governance as a whole. Complexity thinking can augment and link them into a viable environmental governance design. The goal or end state would be a stable human-environment relationship evolving within sustainable boundaries to maintain that balance. The attractors to be addressed include our reliance on carbon-based energy, and our unsustainable use of natural resources, linked to our paradigms of progress and development. The global governance system needs the right spatial and temporal fit while tailoring to very different communities and actors. This means linking the various ideas, processes and institutions into a coherent system that can shape our relationship to the planet. Complexity thinking suggests a number of insights.

While a global treaty or top-down organisation might seem desirable, it seems unreasonable given the highly-fragmented system of agreements we have at present. Instead, complexity thinking asks how they might all work together to address common problems, with some orchestration by a central actor such as UNEP as conductor to set a common direction across environmental governance. All the parts act with considerable autonomy while relating back to the centre. The central authority would have a mandate that cuts across all planetary boundaries but not a regulatory role. It would enable and condition a deeper change in the human-environment relationship, with overarching principles for the planetary boundaries. Its functions would include upholding a set of agreed principles and commitments, an investigating and reporting function for violations of environmental rights and commitments, networking the institutions with common sustainability criteria, and enabling a shift in investments by international financial institutions.

Since our impacts on planetary boundaries intersect and cascade, we need a much faster feedback loop that acts as an early warning and action system, triggering resources and protective action in real time. A scientific assessment of the planet across all major environmental impacts, capturing their interlinkages, would provide a holistic feedback loop of information that generates actions in response. It would include an early warning function based on systemic thresholds and signals of change, an accelerated learning/adaptation function for emerging technologies to experiment with a diversity of options, a co-production function with knowledge generation by all those involved in creating and implementing environmental policies including indigenous, local and practitioner input, and a conditioning role with major actors in energy and industry across sectors included in reporting on their commitments and obligations.

The challenge is for science to break through the present geopolitical and institutional fractures that block collective action. In complex systems, it is necessary to change the underlying rules and create a cascading effect across the system. Collective behaviour change results from an emergent property of the system, not necessarily a set of individual decisions within it. Our reliance on the carbon-based, polluting energy, industry and consumption requires behavioural change on multiple scales including states, industry and consumers. The market will not do this because of subsidies for fossil fuels, barriers to the free distribution of green technology, path dependencies on how and where we produce energy and industry, and inequality. Complexity suggest reform of the financial architecture by shifting financial incentives around carbon and pollution, enabling the free, fast flow of green technology, moving beyond GDP, and addressing poverty and inequality. The Earth4All group has demonstrated that addressing poverty and global inequality would do more for the green transition than almost any other set of actions.

Transformation is required due to a failure of adaptation. Some actors have more power to generate change. This could be a deal between some major countries, a critical mass of smaller states, or middle powers. We need to harness the enormous power of bottom-up approaches, trying to connect them into a counterweight to the resistance of our unsustainable current practices. The important thing is to lock in the transformation, not the institution. Any big bureaucracy or institution resists adaptation to new circumstances, perhaps just adding new layers. We need institutional work, shifting the focus away from institutions towards the efforts to bring about change itself.

The next section of the book is about global security governance, seeing violence as a disease needing an immune system against war. In today’s world of multipolarity and geostrategic competition, military expenditures are expanding, with new escalation around nuclear weapons, emerging technologies and outer space. Violent conflicts are increasing, with many different actors, at multiple levels, driven by climate change and rapidly rising inequalities. War is not only back, but expanding and harder to resolve. There is a mismatch between the complex, interrelated conflict dynamics of today and our global security governance regimes. Our security architecture has been consciously kept complicated, not complex. It is slow, fragile, and unable to respond to new shocks, while failing to address deeper drivers of violence. There is a shift towards securitised, authoritarian forms of governance driving greater levels of inequality and exclusion, moving towards the next round of violent conflicts.

To improve global security governance through complexity thinking, proposals cluster in three main types. The first idea is going big, fixing things at the top, starting with Security Council reform in membership, the veto, and in relation to the General Assembly. Given the low likelihood of a Charter review conference, alternatives might include a Global Resilience Council to address other threats to global security, empowering the General Assembly to take direct action, and bolstering the mandate of the Peacebuilding Commission. The second type is to reconfigure around “clubs” of states with a common purpose in response to geopolitical fracture, taking action first and paving the way for others to follow. The third idea is to go local, either giving greater authority and resources to regional bodies, or through local, bottom-up with initiatives to empower communities to address conflict dynamics themselves. This includes conflict prevention by addressing the deeper causes of conflict such as ethnic divisions, inequality, poverty, and unfair access to resources. The assumption is that state failure and bad governance are at the heart of the problem, but addressing these has always ended in failure.

If we think of violent conflict as part of a complex adaptive system, then we need to ask what the underlying rules and patterns of the system are, how we can gradually alter the gravitational pulls to move away from violence, and constantly adapt to new shocks. The attractors of conflict cross a destructive threshold, de-escalation is reduced, and a positive, escalatory feedback loop increases polarisation and leads to war. Social systems can have a strong attractor towards violence. A global security response should include the following governance functions: a reset button to disrupt endless opposition between countries, a counterbalance towards peace-positive activities or new incentives supporting non-violent actors or processes, disrupting bad information loops and misperceptions with processes that generate solid information that builds trust, and going viral with small changes to entangle peace efforts in the root structures of society. Violence is a disease that can metastasize, infecting our social and political systems.

Complexity thinking can suggest interesting and often counterintuitive directions for global security governance. While strong, inclusive nodes are important within systems, more distributive approaches are also needed, looking at spokes and motion, not hubs. Rather than focussing on the Security Council, security governance mandates and capacities can be distributed outward across the system. This could mean institutional deference with a leading role for other actors addressing non-traditional security threats. The General Assembly could be empowered as a concurrent security actor. Subsidiarity could be turned upside down, with regional and even local actors authorised to take security actions rather than waiting for delegation by the Security Council. Banking for peace would provide a common strategic vision for the World Bank and the UN to align funding flows in the right direction, away from basins of conflict and preventing capture. These ideas reflect convergence where competing authorities are brought around a common set of outcomes and approaches.

In particular, complexity thinking recognises that small changes to rhythms, spaces and routines can have enormous impacts on systemic outcomes. People have a much harder time maintaining violently opposed positions when they are taken out of their usual surroundings, made to walk around, or put in situations where they confront new experiences. The Security Council could be displaced to sites of conflict.

Transparency is important as a feedback loop. Exchange of information reduces the risk of miscalculation, while lack of exchange opens the door to misinformation and disinformation, leading to overreaction in the face of crises and driving towards escalation. To build forums for information exchange and transparency within global governance, a security feedback loop could generate high quality information on security, building consensus on trends and potential risks. There is a need to unblock information flows across the UN system that are often relevant to security risks. Data on food security, climate-related security risks, human rights violations, repression and extremism, all could be part of an early warning system for violent conflict. Then there should be a right to truth, to scientifically-vetted information about issues that matter to our collective security and welfare. Our collective safety requires a common, empirically-founded understanding of risks.

Efforts to fix a problem lead to a flow of money and resources toward centres of attraction, such as an increase in military spending, and peace-building projects to address underlying conflict drivers of inequality and political grievances too often see resources stuck in the centre, captured by a small political elite and the military. An effort is needed to counterbalance the gravitational pull of securitisation. A peacebuilding prism in the UN could reflect resources away from militarisation and towards social spending addressing the deeper causes of violence. Since authoritarian systems drive resources and power to the centre and become entrenched, investing in local elections and political plurality at the sub-national level might have a more disruptive effect. Beyond a UN report on the social and economic impacts of military spending, a demilitarisation conditionality by the World Bank, the IMF and big donors could make reductions in the security sector a priority for support.

Since violence acts like a disease, we need to develop an immunity to war. Communities with a history of violence easily relapse into large-scale conflict. Dictatorship, political violence and repression spread easily. High rates of inequality, political exclusion and a legacy of colonial rule are precursors of violent conflict. Adaptive approaches to peace-building can build a societal immunity or resilience to violent conflict by adopting an evolutionary approach, generating many initiatives together with local authorities, monitoring their effects, and investing in those producing the desired results. At the international level, there may be no room for trial and error, but forums for transparent information flows, shaping the financial landscape away from militarisation, and distributing more decision-making outside of stagnant institutional centres could add flexibility to the governance of security threats. The same flexibility could be added to the rigid state-centred multilateral institutions of security including the UN, turning them into a network to keep options open, allowing new ideas and options to flow in to prevent gridlock. We cannot control the system, but we can help it to avoid getting stuck in the same cycles and entrenched patterns that characterise today’s security discussions. If violence is like a disease, then we need an immune system to produce antibodies to slow its spread and find a good match to beat it back.

Day devotes two chapters to the emerging security and governance issues of cybersecurity and artificial intelligence (AI). On cyberattacks and the problems of endless infection, he describes how vulnerable we have become in our digitally-interconnected society sharing information across a huge number of autonomous but interconnected devices, with the Internet the ultimate complex adaptive system. Cyberattacks can copy and amplify themselves across many computers as positive feedbacks in action, shutting down systems, stealing sensitive contents, and even ransomware demanding payment for access to locked down essential systems. Attacks can grow rapidly and unpredictably, demonstrating complexity. The mainstream concepts and approaches for the global governance of cybersecurity are poorly matched to the type of challenge cyberattacks pose. Traditional defensive tools or state-driven forms of governance cannot prevent them, and creating offensive cyber-capabilities as the best deterrent risks creating positive feedback loops that can move fast.

Complexity thinking suggests networked and distributed approaches that do not disregard multilateral institutions, but instead look to orchestrate and align them. An immune model of cyber-governance will be characterised by adaptation, replicating defensive capacities, inoculating against spread, and building up highly decentralised responses.

The problem is that we live in a porous, deadly, ungovernable cyber-world. Cyberspace is both a global village and virtual battlefield. It combines both physical and non-physical aspects, both the infrastructure of machines, cables, storage capacity and connectivity, which can be regulated, and a social phenomenon of connections between people and institutions in a virtual commons that is neither owned or a global commons. It is a socio-technical system in a network of networks, where evolution, adaptation and ordering of a social system take place on the Internet.

Cyberattacks are difficult to attribute and thus undergoverned, since most actors are in the private sector while international law applies most directly to states. Three common ideas for cybersecurity governance include a cyber treaty against information weapons, a bigger tent in a multistakeholder approach, or the well-defended castle of isolationist national defences that risks fragmenting the Internet itself. Just as the Internet is a complex adaptive system, so should cybersecurity also be systemic, with transparency, not treaties, as disparate organisations and actors agree to principles and practices of transparency, risk-awareness protocols and a shared early warning system for cyberattacks. Fractals, not forums, could involve many institutions drawing on a common set of lessons to create a fractal pattern moving out from them with norm diffusion by the institutions crucial for addressing cyber-risks. We need an immune system, not a fortress.

The metaphors for cyber include the market, where cyberattacks are theft or market inefficiency; the terrain, with cyberspace subject to cyberwarfare requiring military responses; the ecosystem, an evolving, interconnected system of many actors, perhaps requiring protection from invasive species; and public health, requiring computer hygiene to protect from viruses and worms, and suggesting a global immune system to cyberattacks. This would need early warning/detection capacities, a distributed capacity to identify and respond to risks as they emerge, adapting and recalibrating itself to new threats as they emerge, the possibility to quarantine during infections, and a shift of resources and capacities to emerging economies and the Global South where technology growth is already happening. This global immune system approach seems best placed to handle the harsh reality that cyberspace is just as contested, fractured, and fractious as any other arena of politics today.

For the rise of the machines, where artificial intelligence (AI) becomes ungovernable, evolving rapidly towards artificial general intelligence, there are warnings of its threats to the human race. Traditional multilateral tools and institutions cannot address the transformations already underway. The fundamental problem is that AI relies on foundation models trained on large amounts of data to recognise patterns and produce human-like results. Global governance needs to address the development and use of foundation models, but these are opaque, surprising, untestable, easy to misuse, and accelerating. Such a technology does not display its full risk potential until it is deployed, by which time it is too imbedded in our social and economic systems to be meaningfully regulated or scaled back.

The risks of not getting AI governance right are catastrophic if not existential, as it could cause global totalitarianism with social manipulation, autonomous weapons and ubiquitous digital sensors in the hands of a small elite; a great power nuclear war seeing a first-strike advantage; or superhuman capacities not aligned with human wellbeing or values. It could also slowly erode social values and drive global inequalities while accelerating wealth and power. If it acquires human-like characteristics, will it require an ethical response? Our governance response must address these fundamentally new challenges with flexibility for the many ways AI is going to surprise us in a co-evolutionary process.

Four ideas on AI governance include an AI treaty for AI risks, either with strict regulation or key international legal obligations; a global AI regulatory body or scientific advisory body to track AI risks and offer well-grounded policy recommendations; national and regional legislation, at least in the states with most AI development; or norms, ethics and codes of conduct, many of which already exist. Complexity thinking calls for AI governance to be fluid, fit, layered, deep and gravitational. It requires motion, not ossification, since a global AI treaty will be rejected by geopolitical rivals and unsuited to the rapid evolution of AI today. We need to keep conversations moving with a plurality of actors until a new consensus emerges, and especially allow deep ethical reflection on the direction of AI.

Governance of such a complex system must meet five AI-specific challenges: its non-linear change; uncontrollable outcomes as it moves across actors invisibly; proliferation across geographies and domains; concentration of power and resources in a small number of actors; and acceleration as it generates increasingly rapid growth in a wide variety of areas. A good fit for governance would involve an elastic response to surprising developments, capacity to track AI risks at multiple levels, transnational characteristics, safeguards against monopolistic tendencies and concentrations of power, and a scientific analytic capacity. Such governance needs to be careful of positive feedback loops that accelerate trends and concentrate power, and the tendency of AI to concentrate political power and control over populations, whether in states or in the private sector. This requires a negative feedback loop for the risks of AI to society.

Since complex systems tend to be composed of different layers or modules, each may require specific governance processes. The aim of AI governance should be a critical mass, not central control, where enough states cluster around a common set of norms and regulatory approaches, while avoiding the risk of fragmentation. An emerging model of adaptive AI governance would issue reports for assessment against agreed standards, orchestrate global debate, conduct foresight assessments of risks, manage registries of adverse incidents and good practices, promote agreed norms, create a certification toolbox, and direct support to the development of safe and transparent AI programmes. Ultimately, AI might be programmed to govern itself.

In the final chapter, Going back to the future, Adam Day suggests how complexity can help us save the world. He says we are on the brink of extinction, and need to protect the rights of future generations. He describes in detail the problem that we undervalue big future risks, especially catastrophes leading to the collapse of civilisation or our extinction as a species. He describes how to think about the future and to respect future generations. Some of the big ideas on the table include a world government, although no politician would dare use the term, with efforts towards global environmental governance leading the way. Another is a science-policy interface for large scale, systemic risks. A focal point for the future or UN Envoy for Future Generations is already happening.

Complexity thinking can help us to make better decisions about the future by moving beyond resilience as a reaction to crisis, adopting transformative global governance approaches involving a diversity of actors, braking risks before they cascade into catastrophes, all leading to a future oriented global governance architecture like a planetary immune system. We need to avoid the dark side of resilience, including systemic harm and unfair equilibria aiming to maintain the present system, while catching the early signals of systemic collapse already predicted by the Limits to Growth report 50 years ago. This calls for deliberative and designed transformative global governance to address catastrophic and existential challenges, with radical, systemic shifts in deeply held values and beliefs, patterns of social behaviour, and multilevel governance and management regimes, realigning our worldview. Today, a critical mass of people have woken up to existential risks.

Some specific ways that complexity thinking can help to a new design of global governance include diversity in decision-making, braking the system to decomplexify out of trouble, and using systemic design to human-proof global governance. Some characteristics would be to isolate us from thinking only of the present, a gap report on global catastrophic risk, investing in the future, internalising the social costs, recognising the rights of future generations, keep remodelling, and leapfrogging the present. This could take our planetary immune system from complexity thinking to complexity science, with new forms of participation and responsiveness to future risks. Such adaptive governance is forward-facing, a preparation for the future, a positioning of humanity to be prepared for accelerating change.