Introduction

Cascading risks is a well-known perspective in disaster research [3, 80, 95]. The cascades refer to interacting risk factors that amplify the effect of each individual factor if a hazard occurs. Cascades is also a widely used metaphor in ecology [112]. The cascading lens has already been applied to a wider set of risk phenomena, including risks derived from natural disasters being amplified by or amplifying technology risk [49] and even to geopolitical norms [42]. However, that’s only scratching the surface of the cascading risks perceptive.

The cascading effects lens is currently dominated by scholars with a systems theory inclination [3, 20, 36, 53, 75, 101, 106, 113, 134], who explicitly want to deploy a wide systems lens instead of a traditional domain-specific lens. Whilst indeed useful, uniquely using the widest possible systems lens risks missing something essentially important: the core ingredients that go into cascading effects.

In sociology, a dyad denotes a group of two people, the smallest possible social group, who have sustained intimate face-to-face relations over a sufficient amount of time to establish a discernible interacting pattern [13]. In psychology a dyad is used to describe network relationships between significant others or colleagues to understand coping behavior [41, 73]. In epidemiology, a dyad is used to understand risk behaviors affecting community transmission of disease. However, it is also used in political science when analyzing country data [9, 79]. In social network theory, a dyad can also be used to understand the similarity between actors who have not directly interacted [108]. This article extends the dyadic term to look at causal interactions between disruptive forces that indirectly involve humans and which impact humanity’s overall risk level.

As a starting point, the relevance of a set of six conceptually distinct, yet related, disruptive forces, namely science and technology (‘sci-tech’), governance, economics, social dynamics, and health adversity is assumed [123, 124]. In the management literature, this is known as PEST analysis [28]. There is ample scientific support for the interaction between these forces from fields as diverse as the economics of risk [2], disruptive innovation within management studies [24, 114] via multidisciplinary science and technology studies [25, 56, 65, 60], ethics [76], scenario methods in futures research [27, 87, 91, 126], sociology of risk [14, 38, 45] and risk governance in political science [84], to the social determinants of health approach in public health [16, 47], climate change and sustainability studies [35, 64], or existential risk studies [8, 12, 69, 82, 83, 90, 109, 119, 121, 122, 133] which at times attempt to study incredibly complex sets of risks divorced from evolving findings in adjacent scholarly fields–as if existential risks (x-risks) were qualitatively different.

Not all of these forces are included in all types of analysis. In fact, typically a comparative or correlational study will consider only two or three of these forces, although factorial designs permit looking at multiple independent and dependent variables within those domains. The policy impact and risk analysis might be separate pieces (‘deliverables’) of such studies, such as in EU-financed R&D, although impact assessments are used as compulsory policy deliberation instruments during the development of new regulations [37, 99].

Assuming a simple dyadic relationship between each playing out at some point in history would result in 36 combinations. Few studies have described all of these combinations and how they might inform a more structured analysis and mitigation of risk. In this article, research objectives include considering the effects of dyadic hazards but will neither consider more complex triadic or tetradic relationships often called ‘polycrises’ [59], nor will mitigation paths be studied in any detail.

For the record, an example of a four-factor, tetradic risk cascade (a ‘polycrisis’) would be how an economic crisis could lead to the demise of a technology platform which, in a tense geopolitical environment might lead to single- or multi-government collapse which would have significant adverse health effects. The idea behind the present article is that before such complex analyses can be properly conducted, it is necessary to gain a more comprehensive understanding of dyadic risk relationships. The hypothesis is that dyadic risk is more than sufficiently complex to gain an overall understanding of the constituent factors of potential cascades currently at play, that is, to begin cascading effects analysis.

Looking at Fig. 1, the first aspect is naturally that each factor’s interaction with another factor with the same feature represents six of the dyads. A key research question becomes whether each of those is equally relevant or prevalent. Governance is a crucial element affecting risk in society [111]. Additionally, there is the question of whether separate examples for causal relationships are needed or if GovTech should be viewed as 1–1 with TechGov. Such matters are discussed in the methods section. Lastly, note that even though the broader dyadic feature is named Ecological impact, which is more inclusive and representative of broader phenomena including risks to biodiversity, the natural hazards (‘na-’ or ‘nat’) nomenclature from the NaTech tradition is kept to ensure it is distinguishable from economic effects (‘econ’).

Fig. 1
figure 1

The 36 basic dyadic risk mechanisms

Full size image

Methods

The 36 dyadic effect combinations that were selected already at the research design stage were subsequently investigated through a systematic review of the risk literature across a myriad of scholarly fields (notably disaster research, innovation studies, public health, risk science, sociology, and sustainability), using search terms such as cascades, existential risk, risk, impact, interaction effects, natural hazards, and systems, as well as using each of the disruptive factors we use for the six dyadic features under investigation. Literature review is a complex research methodology to execute and the key is to be transparent and follow conscious steps that are motivated by gaps in the field, well explained so as to be reproducible, and which lead to novel findings [77, 96, 115]. I conducted the literature search mainly via online databases, prioritizing Google Scholar and Web of Science, and mostly for journal articles because this search strategy helps to create a more transparent process that can be applied globally [66].

Searches, which took place in December 2023 and January 2024, were first conducted in Google Scholar and Web of Science (WoS) using three search strings: “existential risk”, “systemic risk”, and “cascading risk”, with spot checks using other databases (e.g., ABI Inform/ProQuest, EBSCO/Business Source Premier, JSTOR, MENDELEY, PubMed, ScienceDirect, Scopus, and SpringerLink), but given that the two initial databases had the large majority of relevant content, detailed searches in the latter were abandoned. Given the more limited set of hits, all WoS search results were examined, but results from Google Scholar were further narrowed down by period of publication (2008–2024), language (English), and prioritizing peer-reviewed papers. I made exceptions for a few older references [5, 45]  that were found to be great or unique examples of dyads. The first usage of “risk cascade" outside biology and on systemic risk outside finance was found in the work of German scholar Dirk Helbing in 2012–2013 [54, 55].

Turning to dyads, two specific searches were conducted: (1) “Risk” AND “topical field” (and so on for all 6 framework terms, e.g., sci-tech, economics, governance, social dynamics, ecological impact, health adversity), including variations of spelling and adjacent terms, and (2) “Risk” AND “[scholarly field]” (and so on for all 6 fields under scrutiny (e.g., disaster research, innovation studies, public health, risk science, sociology, and sustainability). Obviously, a larger set of topical or scholarly fields would have yielded additional results, but a cutoff decision was made based on the 80/20 rule considering my initial literature review of the overall risk field. Given the challenging multi-pronged searches needed, I searched using ‘Only in Title’, ‘in Title and abstract’ and only when in doubt or on a hunch, ‘in the text’. An abbreviated list of specific search terms from these searches, including various related terms (5–10 for each category) and search results (‘hits’) is available for interested scholars (see the Supplementary material 1).

For these more extended searches, total Google Scholar hits across these searches were in the millions (n = 8,122,670), so each was rapidly winded down using exclusion criteria such as publication period (2020–2024), and review articles only, which was followed by forward and backward citation searches to identify relevant case studies. From this large set of hits, a matching effort, mapping results to 36 identified dyads (see Fig. 1), resulted in 90 potential abstracts that were reviewed, 30 of which were discarded when looking a bit closer. The 70 remaining papers were subject to full-text eligibility review, and only 10 were discarded. From the 60 remaining papers (including gray literature), a qualitative synthesis of categorized results was created. The peer-reviewed subset of those was 45 articles, in which 205 keywords, 14 keyword clusters, and 4 journal categories were identified (See the flowchart, Fig. 2 and Supplementary material). A total of 12 papers were from Science, Engineering & Technology (SET) journals, 13 from Social Science and Humanities (SOH) journals, 1 from Management journals, and 14 from Interdisciplinary (INT) journals. The average (self-reported or from JCR Clarivate) impact factor across journals was 9.1, with the highest being 64.8 (Nature) and the lowest being 0.58 (Review of Economics De Gruyter).

Fig. 2
figure 2

Dyadic risk—PRISMA review

Full size image

The goal of the literature review was not an exhaustive list of all examples of such occurrences (which is why a reference list of all initial search hits was not produced). Rather the emphasis was on meta-analysis as evidence of the effect of these dyads on the larger phenomena of cascading risks. The analysis proceeded to search for a few good examples of each to form a basic understanding of how useful the joined-up dyadic lens is likely to be in advancing the field going forward. Future research might discover many more occurrences, possibly also further types of findings within each dyad, and at that level of analysis, each could be separate papers.

Even though causal effects were sought, co-causal relationships (such as between TechGov and GovTech) were handled as if they were one and the same as long as the example we found in the literature allowed us to do so. As can be seen, this did not work in all cases. However, if we do treat them that way, we would be left with only 22 dyads (see Fig. 3). The description of dyadic effects could perhaps be subsumed in the same section, but a remaining challenge would be to select the remaining shorthand; do we use TechGov or GovTech? However, the article scopes out all relevant 36 dyadic relationships.

Fig. 3
figure 3

The 22 co-causal dyads

Full size image

Whichever way you handle it, dyadic data analysis presents a number of methodological challenges [86]. Most importantly, dyadic data often violates the assumption of non-independence of variables. In international relations, one study demonstrates how this can lead to overconfidence in the statistical validity of discovered relationships in data on country pairs [39]. Moreover, there may not be enough dyads for sufficient statistical power to create valid data categories. Another challenge might be that each factor or dataset in the dyad may not be on the same analytical level, and as a result, necessitates multilevel analysis.

Additionally, one needs to demonstrate that each of the members of the dyad is distinguishable as separate entities for the purposes of analysis. For that, one needs to show empirically meaningful differences between dyadic factors. One also must show (or assume) that there are no relevant and missed hyperdyadic dependencies, meaning that dyads do not depend upon each other or that there are no systemic effects overriding the dyadic mechanics in importance [29]. Some of these problems can be solved by specific types of regression analysis [79]. Longitudinal data also potentially addresses multilevel or other dyadic challenges.

For the purposes of this article, and because of the large amounts of dyads (36), this paper does not delve further into the methodological challenges of working with dyads as a level of analysis.

Results

I will now present the results of some initial examples of each dyad as found in the literature, with a brief explanation, and as a single data point on the historical trajectory of dyadic risks, denoting the year it was reported. My definitions of each dyad appear in the results section as opposed to in the introduction, because as the empirical work confirmed what each dyadic term would need to include there was considerable refinement needed.

By design, not all examples of risk dyads rise to the level of existential risks which initially were my chief concern. This is appropriate because even midrange risks could become constituent factors of existential risks when occurring as part of a dyadic (triadic or tetradic) cascade of hazards proximate in time and place [125]. Finally, there were literally thousands of articles and examples that could have been used and other scholars will complement this initial selection.

SciTech risk dyads

Sci-tech risks typically impact risks arising from other disruptive factors such as governance, economics, social dynamics, ecological impact, and health adversity, are in turn also impacted by each, and may also interact with them in various ways (Tables 1 and 2).

Table 1 SciTech risk dyads: core concepts
Full size table
Table 2 SciTech risk dyads-examples
Full size table

Note that even though the year reported might show up as relatively recent, no systematic effort was made to find the first time such a dyad was reported and it could have been several decades earlier. Even though most tech-related dyads have been known for a while, the sheer number of articles describing serious concerns about risks stems from the last thirty years or so [6].

Economic risk dyads

Economic risks impact risks arising from other disruptive factors such as sci-tech, governance, social dynamics, ecological impact, and health dynamics, are impacted by each, and may also interact with them in various ways (Tables 3 and 4).

Table 3 Economic risk dyads–core concepts
Full size table
Table 4 Economic risk dyads-examples
Full size table

I note that the impact of dyads is often multifaceted and can initially be negative and then turn to a positive, both as regards risk and innovation. This relates to the complexity of analyzing dyadic relationships, as hinted at before.

Governance risk dyads

Risks derived from governance impact risks arising from other disruptive factors such as sci-tech, economics, social dynamics, ecological impact, and health dynamics, are impacted by each, and may also interact with them in various ways. Governance is here widely understood as attempts to control, influence, or structure rules, norms, and actions through policy measures, regulatory efforts, and rule-making behavior across public, private, and nongovernmental institutions–involving both centralized, decentralized, and mixed approaches (Tables 5 and 6).

Table 5 Governance risk dyads-core concepts
Full size table
Table 6 Governance risk dyads-examples
Full size table

Broadly, the governmental dyadic risk effects were reported earlier than other areas, perhaps because risk is a relatively obvious concern when it comes to the government’s core responsibilities.

Social dynamics risk dyads

Risks derived from social dynamics, broadly understood as group behavior, relationships, and interactions, impact risks arising from other disruptive factors such as sci-tech, governance, economics, ecological impact, and health adversity, are impacted by each, and may also interact with them in various ways (Tables 7 and 8).

Table 7 Social dynamics risk dyads-core concepts
Full size table
Table 8 Social dynamics risk dyads-examples
Full size table

Social dynamics surrounding technology have been analyzed almost since the beginning of industrialization, beginning with the founding fathers of the discipline of sociology. More recently, the field of science and technology studies (STS) has taken up the challenge [107], and in the last decade, a growing body of findings from complex science and systems perspectives has started to shape the field [75].

Ecological risk dyads

Natural hazards impact risks arising from other disruptive factors such as sci-tech, governance, economics, social dynamics, and health adversity, are impacted by each, and may also interact with them in various ways. Natural hazards are events (geological hazards, hydrological meteorological hazards, or biological), such as avalanches, blizzards, cold waves, droughts, earthquakes, fires, floods, heat waves, hurricanes, landslides, storms, tornadoes, tsunamis, volcanoes, or wildfires, with a likely negative impact, potentially killing thousands of people and destroying billions of dollars of habitat each year.

Ecological risks are wider than natural hazard risks because they include risks potentially with no immediate event associated with it (air contaminants, biodiversity collapse, harmful chemicals, infectious diseases, pandemics, pesticides, plant and animal invasions, toxic waste, water-borne diseases) but rather may stretch out in time with a more insidious impact (Tables 9 and 10).

Table 9 Ecological risk dyads-core concepts
Full size table
Table 10 Ecological risk dyads-examples
Full size table

It took a significant amount of time for NaTech risks, risks associated with the combined effects of natural hazards and releases from industrial facilities, to be identified and properly discussed [106]. It is now important to broaden the scope of analysis to wider science and technology-related factors, phenomena, or platforms, whilst keeping a focus on the negative impact of industrialization and industrial infrastructure on nature.

Health adversity risk dyads

Health adversity risks impact risks arising from other disruptive factors such as sci-tech, governance, economics, social dynamics, and ecological impact, are impacted by each, and may also interact with them in various ways (Tables 11 and 12).

Table 11 Health adversity risk dyads: core concepts
Full size table
Table 12 Health adversity risk dyads: examples
Full size table

The risks of health adversity have long been a subject of study in the field of public health. With the increased potential for innovation using emerging technology also comes greater risks, which have so far received far less attention. This is strange, given that health consequences are the ultimate impact that matters most to humanity. It also matters a lot to distinct and often disadvantaged social groups. Lastly, health adversity, as well as its corollary, perceived wellness, is a powerful individual concern and a motivating force.

Discussion

The alignment of dyads

A basic clustering of the papers used to exemplify dyads reveals 14 clusters with cross-cutting issues such as finance, environment, governance, health, social impact, and technology, closely matching our input framework’s six dimensions (see Fig. 4). Whilst not particularly surprising, this also validates the use of that framework. Despite the complexity of dyads, they align along commonly considered dimensions. What’s lacking is putting these insights together to understand how they can cause cascading interactions or possibly can be mitigated before they do.

Fig. 4
figure 4

Keyword clusters

Full size image

The complexity of dyads

The paper sought to illustrate how dyadic risk is more than sufficiently complex to gain an overall understanding of the constituent factors of potential cascades. Looking across dyads, the phenomena described are intricate yet addressable within their field of study. To what extent does the analysis of these 36 combinations inform a more structured analysis and mitigation of risk? It turns out that it is difficult in most of these 36 dyads to separate out the causal relationships from the analysis of the dyad. In other words, are we looking at one factor’s impact on the other or is there a systemic interaction within the dyad that is responsible for the impact? This is relevant for assigning causality (and risk responsibility) but also for understanding the nature of emerging risks.

Logically, if A and B represent each of the dyadic factors in a pair, the basic causal options are:

A—> B (A directly causes B effects) or B—> A (B directly causes A effects). However, we must also account for co-causality, meaning that A and B—> A (A and B jointly cause A effects) or A and B—> B (A and B jointly cause B effects), or A and B—> Both A effects and B effects. In reality, of course, both A and B are embedded in far more complex causality chains where there are intervening variables (A—> C—> B) or even spurious effects with an unknown factor causing A B or both A and B (C—> A and/or B) or even pure coincidence because so many factors could have caused a change in either A or B, for which there are now formal coincidence analysis methods [11]. Systemic risks in complex coupled systems are not obvious because causality is obfuscated by a multitude of intermediary factors (C, D […] to Z—> A and/or B) whose relationships (how C relates to D, etc.) may lack predictability [106].

Although we have not assessed the causality implications of all the examples provided in our sample of 36 dyads, it would probably be fair to say that correlation is more prevalent than causality. I should be quick to point out that some of the problems noted with dyadic analysis are not necessarily a problem with the original studies cited because they were discussed and appropriately handled by those researchers. Dyadic causality only becomes a larger issue when we attempt to use those same studies for a multi-dimensional cascading risk analysis.

If we even just consider one of the two factors in a given dyad (A and B), each of the 36 pairs had quite complex relationships even within each factor, with potential relationships between thousands of variables. If we consider EconSoc or TechSoc, the relationships between the economy and society, for example, are manifold and are the main preoccupation of the entire field of economic sociology. Both the economy of risk and social dynamics of risk are complex fields subject to varied scientific analysis (from across the social sciences and beyond), as we saw when looking at E2 and S2.

Looking across a few dyads, it is not clear that we can analytically separate economic from governance effects. Indeed, the field of political economy, which involves the study of how economic theories such as capitalism or communism play out in the real world was a reaction to this. Adam Smith, Thomas Malthus, David Ricardo, and later, Keynes and others, criticized mercantilism’s limited understanding of economic nationalism, the idea of a zero-sum game of trade where some benefit at the expense of others. To understand the economy, they argued, we need to understand politics and the economy as one system [7].

We should be careful in interpreting surface-level observations about dyadic risk relationships, and making inferences about wider cascades, both for methodological and practical reasons. Each example given above needs to be examined closely to figure out what the timelines were and how each dyad is likely to be affected by the passage of time. However, for historical examples, we can already do so. A recent study of the history of risk-tracking semantic patterns over a period of 150 years found that the risk discourse reflects changes in terms of the type of threats (away from wars and toward chronic disease) as well as changes in the potential to forecast, prevent, or mitigate them [74]. How much of these effects only represent increased confidence within the growing risk literature as opposed to real ability to predict future risk, is another question.

Dyads as building blocks for causal analysis

Broadly, then what have we learned about dyadic relationships of risk? We can certainly now in an easier manner, using the shorthand provided, cursorily describe a huge amount of risk data and some macro effects that govern their relationship. This could be useful for policymakers, in risk assessment, or in communicating broad areas of risk to the wider public. We have also provided a basic building block for more complex factor analysis that might lead to more advanced cascading effects models (of pentads, hexads, heptads, octads, and onwards).

Having said that, the number of factors escalates with each level of interaction (6 to the power of 2 is 36, but 6 to the power of 3 is 216). Thoroughly describing the 216 triadic risk relationships in the manner we have done here for 36 dyads, for example, would likely be out of scope for a scientific paper. Imagine that we had to consider 1000 variables within each factor (A1-A1000) and how they interrelate even before the dyadic and triadic analysis. It is highly possible that the future of cascading risk analysis lies in AI-assisted factor analysis to complement advanced statistical techniques such as multivariable logistic regression. So far, this type of approach has been usefully applied to cybersecurity, where the ability to quickly detect, analyze, and respond to threats is paramount [22]. However, we would rapidly reach the limit for human interpretability of the results, a problem all too common with AI-powered analysis. Also, we would likely need quantum computers to analyze the results.

Economists have, of course, with the crude powers of statistically empowered economic analysis, and notions of limitless growth, had the hegemony within most governmental administrations for the latter part of the century [104]. Adding a second factor to the mix, tech, the techno-economic paradigm has dominated the last few decades [94]. Modernists would make a similar argument for focusing on sci-tech innovation and impact [132]. Is that role now challenged as powerful methods and approaches are also increasingly available in other domains? The impact and explanatory powers of other domains of inquiry (across the social sciences) are becoming clearer. Given the broad scope of impacts upon lifestyle, quality of life, longevity, suffering, inequality, life and death, and more, we might make a similar argument for health innovation [43] or health adversity effects [92]. However, the fact that disciplines do already meaningfully engage with risks that, when interacting, constitute cascading risks is comforting. Stepping up that effort might be possible from within scholarly fields and may not require as much of a step change in existential risk studies as that discipline assumes [90]. On the other hand, grasping existential risks requires a deeper engagement with existing scholarly findings in the six fields tracked here than is currently the common practice in the x-risk literature.

Dyadic governance

Given the importance of gaining confidence in society’s ability to manage risk, getting a better grasp of the six core governance effects (GovTech, GovEcon, G2, GovSoc, GovNat, and GovHealth) is certainly crucial. GovNat effects, arguably one of the great contradictions of our time, maybe self-inflicted but are hard to counteract given that they are steeped in power structures and long-term cycles. At this point, industrialization’s effect on global warming has been analyzed to death, and the effects are well known [64], but the mitigations are ineffective, expensive, and politically difficult to implement given that they would deteriorate industrial conditions affecting the world’s largest companies even beyond industrial actors. Moreover, they would reshape the destinies of powerful nation-states and economic actors such as the US, the EU, China, and India. Negotiating solutions will certainly occupy the world’s elite for decades, potentially centuries to come. However, would an even more useful path for researchers be to start mapping out more recently emerging GovNat effects so that we could potentially forecast, prevent, or mitigate some of those? Governmental influence upon biodiversity would seem to be the next frontier and the amount of work ongoing is far inferior to the effort currently going into climate change engagement. Mapping these dyads indicates that governance-related dyads are crucial because they have the potential to shape all other dyads. This is not to say the task is easy–there is still complexity inherent in learning the right lessons and establishing and maintaining governance across this wide scope of topics and fields.

Scholarly engagement complementing interdisciplinary outlooks

It is currently common to argue in favor of the urgent need for inter- or transdisciplinarity and systemic perspectives on risk [75]. This paper is not an argument against that but provides a supplementary motivation: disciplinary engagement with risk could also be pivotal. The good thing is that the present study found ample examples of engaging with dyadic risk across disaster research, innovation studies, public health, risk science, social sciences, and sustainability studies. Mapping the impact against the likelihood of individual risks is a true and tested approach in engineering-driven risk management of large-scale industrial endeavors. Establishing true values for these risks is no trivial task. Even delineating which factors need to be included in the matrix is tough to do, even if I have simplified that process by assessing 36 dyadic risk pairs. However, there can be no path to mitigation without having at least a cursory understanding of which factors might matter, which might impact each other, and how large the combined effects might be. Part of this work entails scholarly risk studies within the highlighted disciplines. Part of it consists of existential risk studies engaging more deeply with the findings of those disciplines. Very little cross-referencing between these literatures was found in our sample to indicate that this is the case.

Conclusion

In this paper, through a systematic literature review, I analyzed 36 dyadic risk mechanisms that emerged from a juxtaposition of pairs of two (‘dyadic features’ of risk) derived from six previously identified disruptive forces: sci-tech, governance, economics, social dynamics, ecological impact, and health adversity [123, 124]. Future research is needed to determine whether additional forces, institutions, sectors, or spheres, such as culture, media, industry, education, the nonprofit sector, religion, space, or others also deserve their own shorthand acronym and inclusion in a comprehensive dyadic risk mechanism review.

Even though there is a large body of research that has touched on the themes covered, no previous studies have covered all of them systematically. Present findings show how complex risk relationships can be, even at their most embryonic stages. In the real world, risk rarely shows up as a phenomenon only impacted by two main factors. As a caveat, most of the dyads illustrated here are so complex that they definitely qualify as systems themselves. Even if a larger systems perspective might be the ideal tool in the risk analysis toolkit it would seem that a lot of the preliminary work of analyzing simpler phenomena and how they interconnect has to be carried out simultaneously, or even before attempting such a feat. Our current understanding of systemic risks is so limited, and the need to go deeper so pressing, that delving into such dyads seems a necessary step towards tackling cascading risks more generally and expansively, and certainly before attempting to thoroughly describe existential risks that would entail a plethora of interacting factors. There can be power in simplicity, too.

Another benefit of the dyadic approach is that it complements the systems theoretical approaches taken by the disaster research community as well as the existential risk community with the notion that existing scientific knowledge and scholarship increasingly need to take on board risk mechanisms instead of attempting to simply separate them into distinct fields of disaster risk and existential risk. If this could happen, the available scientific talent to address concerning risk challenges would multiply. Both the disaster research lens, the futures research lens, and the existential risk lens are missing elements that may be fundamentally important for the scientific validity and political legitimacy of necessary foresight, innovation, resilience, and mitigation efforts surrounding the totality of such risks. However, risk engagement cannot simply be a side deliverable, it has to be embedded from the start.

The research question in this paper was whether each of the 36 dyadic risk mechanisms is equally relevant or prevalent. Naming them and developing a concept sentence for each is the first research contribution made in the present article. I determined that six of the most dire cascading risk mechanisms were T2, NaEcon, GovNat, G2, NaHealth, and TechGov effects, and each deserves considerable scrutiny in the time to come.

Moreover, the argument could be made that the six governance-related mechanisms are of primary importance at least from the perspective of risk mitigation. Among those, the GovNat effects such as governmental (non-) regulation of climate change, are the most impactful at present. However, those effects are directly related to the TechGov and GovTech effects because government R&D often finances platform technology [34] and definitely has the potential and responsibility to regulate impactful platforms once formed.

Given the progress and increased interaction of emerging technologies and platforms, there is also the argument that T2, T3, and T4 effects should receive more attention in the years to come. That will require expertise across those technologies and platforms and talent with those skills and also willing to dedicate themselves to assessing risk as opposed to, or in addition to innovation will be in short supply.

Future research should definitely focus on creating a more exhaustive list of research and case studies underpinning each of the 36 dyads sketched in the present paper, including establishing the year when each dyad was first reported. Moreover, future research should also attempt to further simplify this framework, or if found useful, expand the number of core factors, and should aim to discover whether and in what contexts a few of the core mechanisms determine the bulk of the risk. This could be a subject of historical significance. But in a rapidly changing world, we also need to beware of making too many assumptions based on the past. Dyadic risk mechanisms seem to have been important factors in shaping our past. Armed with that insight, we can perhaps soon make a more informed decision about whether they should also determine our future.