By Caleb Peppiatt
Suffering-focused ethics is a set of moral views that place particular importance on the prevention and mitigation of suffering. This includes the prevention of “risks of astronomical suffering” that might be brought about due to humanity’s efforts to colonise space (Kovic, 2021; Baumann, 2022).
Previous debates in space ethics have rarely placed particular significance on the suffering of future populations (Brennan and Lo, 2021; Owe, 2023).
While consequentialists have discussed space colonisation, it has largely been to advocate for it, and they have often assumed a moral symmetry between suffering and positive phenomenal experiences (Greaves and MacAskill, 2025).
Suffering-focused ethics remains a somewhat neglected perspective in debates on space colonisation. My book, The Case Against The Stars, fills this gap by systematically assessing suffering risks in non-terrestrial contexts. This essay provides a condensed statement of its argument that we ought to refrain from initiating effectively irreversible colonisation processes unless and until robust suffering-mitigating constraints are demonstrated.
Contents
I – Risks of Astronomical Suffering in Space
II – A Precautionary Argument
III – Conclusion
I – Risks of Astronomical Suffering in Space
First, consider a select set of suffering-risks associated with space colonisation.
Directed panspermia refers to the intentional seeding of life to other planets, an activity which will soon plausibly become technologically feasible for a large number of incentivised actors (Soryl and Sandberg, 2025). If successful, sending probes to other target planets could kickstart evolutionary processes on a vast number of star systems, extending Darwinian struggles deep into the future. This would likely generate vast quantities of severe suffering, given that R-selection, scarcity, and antagonism would be expected to occur in nearly all evolutionary processes (O’Brien, 2021, 95; Sivula, 2022, 180, 193).
Terraforming refers to the modification of environments to enable surface habitability, requiring the continued maintenance of atmospheric conditions (Turyshev, 2026). Terraforming could enable vast numbers of biospheres to be created, thereby multiplying wild-animal suffering. This risk is distinct from directed panspermia: terraforming plausibly requires active maintenance by colonists to stabilise surface habitability, such that the duration of suffering is bounded by the decisions of colonists; similarly, if colonists co-habit terraformed environments alongside wild animals, they might intervene to mitigate wild animal suffering – unlike with panspermia if target planets are too distant for human travel; finally, forward and backward contamination between terraformed planets could plausibly result in mass mortality and mass morbidity incidents (Hamilton, 2022).
Farming practices extend the exploitation of captive animals to non-terrestrial environments. This includes animals living in bio-regenerative life-support systems as would be deployed in shuttles and orbital settlements (De Micco et al, 2023; Shevtsov et al, 2023, 4), as well as intensive farming practices on lunar and planetary surfaces (Przybyla, 2021; Ulrich, 2025). Given the potential quantity of immense suffering incidentally produced in terrestrial factory farming and aquaculture, and the limited concern humanity currently displays toward non-human animals, plausible magnitudes for suffering produced in non-terrestrial settlements suggest that there is a potential for astronomically amounts of severe suffering.
Totalitarian political arrangements might plausibly thrive in isolated colonies and generation ships (Cockell, 2022). This is partly due to the immense distances separating colonies undermining oversight and normative pressure from other colonies, the lack of exit options for exploited human colonists, and incentives favouring surveillance and control in order to maintain life-support systems.
Warfare might plausibly break out between colonies and space-faring civilisations due to normative and memetic divergence (Torres, 2018, 75-7), preemptive attempts to maintain security (Jebari and Olsson-Yaouzis, 2018), and malevolent actors (Althaus and Baumann, 2020). Warfare could be enabled by technologically feasible, yet indiscriminate and imprecise means, including kinetic bombardment (Deudney, 2020, 180, 368-72), the redirection of asteroids (Daly et al, 2023), blockades of imports, and strikes against life-support systems.
If space colonisation involves artificial minds, this produces another set of risks. Artificial minds might provide cheap and duplicable sources of labour, with significant incentives existing to exploit such sentient beings (Schwitzgebel and Garza, 2015; Bradley and Saad, 2026, 6). Likewise, exploiting sentient beings within simulations may be plausibly incentivised in order to predict behaviour and acquire information (Bostrom, 2003b). Artificial minds propelled in Von Neumann probes might additionally be able to colonise space at rapid speeds (Armstrong and Sandberg, 2013, 12). If technically possible, the competitive expansion of artificial minds would plausibly select against welfare (Hanson, 1998; Buhler, 2023).
Second, consider shared features underlying these risks.
Features of space amplify the magnitude of suffering associated with each risk: the sheer number of habitats multiplies independent loci of wild-animal suffering; the extension of such risks into deep time allows suffering to occur beyond terrestrial timescales; the immense scale of resources allows novel sentient populations to be created at scale.
These risks are also highly difficult to reverse. Sending out microbial payloads or Von Neumann probes are processes that most likely cannot be recalled once initiated, and preventing such actions from occurring is highly difficult (Soryl and Sandberg, 2025, 400). Other activities, such as establishing highly-capable AI systems that maintain life-support systems, might entrench governance structures and values prematurely. Interstellar distances undermine the capacity of actors to monitor each others’ behaviour and counteract the imposition of totalitarian control over colonies.
Further, such risks are comparatively technologically feasible – technologies enabling these processes are differentially feasible compared to relevant suffering-mitigating technologies. For instance, farming systems that harvest sentient organisms are more feasible than farming systems which do not, such as those that rely upon cost-competitive cultured meat (Humbird, 2021). Directed panspermia efforts are more feasible than efforts to create biospheres full of sentient organisms that do not suffer. Suggested methods to hedonically engineer organisms such that they do not experience severe suffering (for instance, Pearce, 2015) are themselves dual-use, and might be abused to create even worse suffering.
Finally, many of these risks have a non-negligible probability of occurring. Attempting to precisely estimate the probabilities and expected magnitudes of each risk is not appropriate. However, one can still be justified in holding that these risks have a non-negligible probability of occurring: the risk landscape is disjunctive, as each risk can be realised by a multiplicity of actors across space and time. The disjunctive probability that at least one actor – across a vast number of potential actors over billions of years and vast volumes of space, would initiate such irreversible processes is rendered non-trivial by structural pressures, incidental incentives, and the difficulty of establishing robust governance mechanisms.
II – A Precautionary Argument
The following precautionary argument does not require strong suffering-focused intuitions or assumptions, and instead uses a form of precautionary reasoning that is commonly applied in environmental ethics.
It applies most forcefully for colonisation activities that are hard to reverse. This includes the initiation of self-perpetuating processes which propagate without further authorisation (such as directed panspermia, Von Neumann probes, and establishing colonies capable for autonomously launching further expansion); the entrenchment of particular values or governance mechanisms (such as establishing AI-enabled surveillance and enforcement systems); and the creation populations that cannot be feasibly governed at scale.
P1: For any action A, if
- A is effectively irreversible; and
- A carries a non-negligible risk of astronomical suffering; and
- That astronomical suffering is not reliably preventable if A is initiated under present conditions; and
- A can be postponed without forgoing benefits comparable to the suffering instantiated through A (refraining from A preserves the possibility of performing A under better future conditions),
Then one ought not perform A now.
P2: The initiation of irreversible processes of space colonisation – such as directed panspermia, establishing self-sufficient, autonomous colonies, and the creation of vast biological and digital populations – satisfies (a)-(d).
Therefore,
C: One ought not initiate irreversible processes of space colonisation now.
Premise 1 restates one of the most sophisticated and defensible precautionary principles in terms of astronomical suffering rather than “catastrophic disvalue” (Gardiner, 2006).
Conditions (a) through (c) of Premise 2 were implicitly defended in section I.
Condition (d) of Premise 2, that refraining from initiating space colonisation processes does not forgo “comparable benefits,” depends upon which suffering-focused theory is held. Under certain lexical views, there are no “comparable benefits” that could possibly justify the instantiation of severe suffering. Other positions might theoretically allow positive goods to “outweigh” severe suffering, and instead weigh lost hedonic pleasure as of less ethical significance. One might simply contend that comparable benefits would not be forgone by postponing space colonisation, which even symmetric utilitarians such as Nick Bostrom hold.
A similar argument can be made in terms of justificatory burdens:
P1: Those who impose novel, irreversible risks on others, including future populations, ought to bear a justificatory burden for how risks will be prevented.
P2: Where the disvalue at stake is catastrophic and irreversible, the justificatory burden that the proponent must meet is unusually high.
P3: In the case of space colonisation activities proposed by longtermists, novel, irreversible risks are imposed on others, and the disvalue at stake is catastrophic and irreversible.
Therefore,
C: The proponents of space colonisation activities ought to bear a justificatory burden, and this burden is unusually high.
For a more complete articulation and defense of these arguments, see The Case Against The Stars.
III – Conclusion
Suffering-focused ethical perspectives have not yet contributed to debates in space ethics as much as they ought to, particularly given the vast sizes of potential future populations. This report has provided a brief overview of risks of astronomical suffering associated with space colonisation, and has provided a precautionary argument in order to defend the following conclusion:
Suffering-focused ethical perspectives recommend refraining from initiating irreversible space colonisation activities unless and until robust suffering-mitigating measures are demonstrated.
Various theories of value and normative commitments differ in how they assess the value of colonising space (Peppiatt, 2026b). However, both suffering-focused intuitions and precautionary reasoning favours restraint. If one views the mitigation of suffering as of enormous moral significance, then this conclusion provides a strong reason to be cautious over proposals to expand across the stars.