Some thoughts on how an effective altruistic organisation might think about prioritising different types of S-risks. Quick primer here.
Agential s-risks are in themselves a form of incidental s-risks—not wholly, as agential s-risks involve some practically random/uncontrollable and individual aspects, but this essential subordination is substantial enough that these Russian doll dynamics should inform strategic priorities. Here are some examples of this claim.
By perpetuating and promoting dysfunctional societal dynamics characterised by fear, scarcity, inequality and misery, a dominant/institutionalised ideology simply aiming to promote economic growth based on a capitalist model may be the perfect breeding ground for directly human-agential s-risks (as opposed to indirect agential risks associated with technology, which I will briefly touch on later).
This type of incidental context can increase agential s-risks:
- by encouraging sadism (e.g. children with traumatic childhoods, without protection or remedies from appropriate social structures, are more likely to abuse animals in the commonly accepted sense of the term);
- by promoting sociopathic behaviour (e.g., poverty is widely considered a determinant in criminal behaviour in later life). Early adverse experiences impair the development of parts of the brain associated with decision-making and emotion, for instance. It seems likely that these experiences are most likely to occur in societies where a focus on economic value-creation normalises the deprivation of some segments of the population. It should be noted that relative disadvantage, and a competitive rather that cooperative societal framework, are far more relevant in assessing risk creation than low but fairly even levels of economic abundance and access to opportunities, both for parents and for children;
- by promoting identity politics and a perceived imperative to compete, individually but also collectively, for status, resources, and security. This can lead to tribalism and retributivism, which can both be partly motivated by efforts to increase community cohesion through the deliberate infliction of violence onto others.
Relatedly, societies structured around the extraction of resources place a strong emphasis on militarism, which creates a natural path to positions of significant agency for malevolent actors, since psycho-pathological traits such as lack of empathy, boldness, and violence are rewarded with increased access to power, and therefore opportunities to cause suffering, e.g. among conquered populations during imperialist campaigns or occupation/colonisation, but also in military rising stars’ future careers as statesmen/dictators. Militarism in support of constant expansion and value- and market-creation is essential to our current societal paradigm and therefore represents an incidental s-risk—both because of the suffering involved in armed struggles and because of the opportunity it represents for malevolent agents to increase their agency in material and temporal scope. This is how dictator Francisco Franco came to power, for instance.
To provide a sense of the intricate and weighty impact of incidental determinants on agential s-risks, it is also useful to consider the impact of political inertia towards, for instance, mass exposure of children to manganese and lead despite these heavy metals’ apparent role in criminal and violent behaviour. Such practices, sometimes as passive in nature as failing to prioritise issues such as human welfare over perceived economic constraints, are endemic s-risk factors, incidental in that they occur at the societal level and not for malevolent reasons, which may bring about the emergence of wilfully nefarious individuals who could become perpetrators/instigators/designers of mass suffering.
Continuing with the present theme of considering incidental aspects of agential s-risks from a critical political economy/sociology perspective, school shooters, a common type of malevolent agent in some capitalist societies, subject numerous communities, across several generations, to significant suffering relative to the ease of perpetrating their acts of violence and Western societies’ vulnerability to such attacks. Might society-wide emphasis on individualism and competition, the toleration with which marketing fuels frustrations and insecurities, ready access to powerful weapons, and knowledge of the potential to cause anguish, be equally decisive in the occurrence of school shootings as the perpetrator’s psychological predisposition to act malevolently due to their genetic makeup, for instance?
Of course, these small-scale instances of deliberately inflected torment are not the stuff of s-risks—the point of these examples is simply to emphasise the role of societal/incidental factors in the intentional exaction of suffering, whatever the scale of the affliction caused.
Non-human agential s-risks could also arise, at least in part, due to incidental hazards. If artificial sentient agents come to be created due to lack of precautionary regulations as artificial intelligence develops due to the primacy of market forces over other societal priorities, and if these agents are capable of wilfully imparting suffering along the lines of human cruelty, this constitutes both an agential s-risk—and thus one potentially fraught with impulsiveness and irrationality, or with complete lack of empathy or concern for others’ interests in avoiding suffering—and an incidental s-risk with the potential to generate suffering on a scale and timeframe matching the reach of the possibly highly intelligent, and therefore resourceful and persistent, agent causing it.
As exemplified above, incidental s-risks encompass their own hazards, but also key aspects of agential s-risks. Therefore, unless agential s-risks can be considered substantial causes of incidental s-risks, or are likely to exceed in scope the impact of incidental s-risks, it seems likely that incidental s-risks are a greater cause for concern than agential risks from the perspective of individual instigators.
However, for incidental s-risks to be the most effective area of focus for a research- or policy-oriented organisation, they have to be the most tractable. While agential s-risks can arguably only be tackled through the implementation of safeguards against individuals at risk of deliberately seeking to exact torment on enormous numbers of subjects, it is also the case that incidental s-risks present higher tractability since they can be addressed through public policy affecting social causes of agential infliction of suffering.
For instance, on the one hand, ending factory farming (incidental s-risk) and reforming criminal justice (both a potential incidental s-risk if this blueprint persists as humans multiply, colonise space, and so on, and a determinant of agential s-risk) can both be done through policy. On the other hand, monitoring every agent throughout their lives—or even screening their genetic material before birth, perhaps, to control individuals predisposed to psychopathy—ensure societal triggers, and aggravating hazards remain out of their reach, or that fail-safes are in place to dampen deliberate efforts to mete out suffering, is all at once less practical, less self-sustaining, less reliable and possibly more ethically problematic (e.g. by requiring invasive scrutiny over agents and/or limitations to important rights and freedoms, and by diverting resources that could otherwise be used towards promoting social goods). As applying concern with suffering to practical situations does not happen in a vacuum, it is noteworthy that focusing on incidental s-risks, or perhaps, more aptly, incidental aspects of s-risks, may also improve overall positive welfare by the same token.
To briefly explore natural s-risks, evidence—notably the state of academic progress—suggests that at least some natural s-risks require more scientific research before they can actually benefit from the scrutiny of researchers providing strategic ideas.
Natural s-risks are undoubtedly neglected, and among natural s-risks, the effectiveness of tackling wild animal suffering is now becoming more promising thanks to advances in welfare biology and related disciplines, and possibly, a gradual expansion of humanity’s moral circle. Given the scale of the anguish involved in wild animal lives, and humanity’s nascent purchase on practical, viable solutions to causes of wild animal suffering, it seems reasonable that researchers should devote some resources to considering these problems and strategies.
However, for a researcher/organisation external to directly relevant scientific fields, it would likely be comparatively ineffective to expend a high proportion of resources on this very important issue. This is because more technical ‘disentanglement’ still needs to be provided by animal welfare specialists and other natural science researchers to inform organisations seeking to design advanced approaches to reducing suffering among wild animals based on solid knowledge on sentience, ecosystems, animal physiology and adaptation, and the like. It appears crucial to avoid the potential for unintended consequences and path-dependencies with very poor outcomes caused only by lack of understanding when there are more secure areas of focus, such as incidental s-risks.
Moreover, although the strength of the following argument deserves to be evaluated with great care, some animal welfare stakeholders claim that wild animals living in their natural environments are faced with types of suffering they have evolved to cope with, as opposed to captive wild animals (whose welfare is arguably best considered outside of natural s-risks, as their suffering is at least partly manmade) and domesticated animals subjected to intensive exploitation or experimental procedures, for instance. This argument leads to the conclusion whereby free-living wild animal suffering may be less pressing than other forms of animal suffering.
A basic flaw with this argument, of course, is that wild animals’ ‘natural lives’ are increasingly impacted by human activity, which challenges their very chance of suffering only in ways they have evolved to cope with better than forms of suffering provoked by human-driven habitat destruction, stress associated with human interaction and similar anthropic factors. Clearly, wild animal suffering is also a by-product of human activity, and therefore incidental as well as natural.
On the whole, it seems plausible that tackling incidental (and thereby also agential) s-risks through policy is, currently at least, the most tractable area for research on reducing suffering, as it relies on fairly established, non-negligibly effective structures and processes—legal reform, public education, and delivery of public goods/allocation of resources, individual and collective interplay with social contract, harnessing of market dynamics, and so on. Natural s-risks, on the other hand, can only be directly mitigated in proportion with our limited grasp of and agency ‘over’ nature. Nature is a set of complex systems that were not designed to respond to human agency on a significant scale.
Therefore, at present, at the dawn of effective scientific progress, natural s-risks are only moderately tractable. This is especially true when it comes to potential very-low-level suffering at the microscopic/particle scale, a realm that currently requires great resource expenditure to explore adequately based on the sheer number of potentially suffering particle. I am currently sceptical as to how concerned we should be over the potential for non-sentient entities to suffer, because:
- it appears likely that suffering emerged due to evolutionary imperatives that do not seem likely to affect physical particles,
- the urgency in dealing with the vast amounts of certain suffering endured by human and non-human animals in the present offers plenty of room for reducing suffering effectively—if we have to prioritise, which the state of investment in preventing mass suffering suggests is the case, and
- physicists probably still have a significant amount of work to do before humanity can routinely place shock absorbers onto the electrons of 1080 atoms or do whatever else reducing suffering in such particles would look like. This suggests it may, for now, be worth prioritising the reduction of the very real, and already quasi-astronomical amount of sentient suffering present all around us.
To sum up, I claim that due to their multifaceted forms of impact on all s-risks, built-in potential responsiveness to strategic steering, and positive effects associated with tending to social goods in working to prevent s-risks, incidental s-risks currently constitute a judicious area for organisations seeking to prevent mass suffering to prioritise working on.
The Sound And The Theory 