Space governance is important, tractable and neglected


I ar­gue that space gov­er­nance has been over­looked as a po­ten­tially promis­ing cause area for longter­mist effec­tive al­tru­ists. While many un­cer­tain­ties re­main, there is a rea­son­ably strong case that such work is im­por­tant, time-sen­si­tive, tractable and ne­glected, and should there­fore be part of the longter­mist EA port­fo­lio.

I also sug­gest crite­ria for what good space gov­er­nance should look like, and out­line pos­si­ble di­rec­tions for fur­ther work on the topic.

What is space gov­er­nance?

It’s plau­si­ble that hu­mans, or their suc­ces­sors, will even­tu­ally be able to colon­ise space. There are already var­i­ous Mars mis­sions, and fu­ture tech­nolog­i­cal ad­vances might make large-scale colon­i­sa­tion eco­nom­i­cally fea­si­ble.

Space gov­er­nance en­com­passes the laws, rules, norms and in­sti­tu­tions that struc­ture in­ter­ac­tions in space, as well as mechanisms that are used to es­tab­lish and en­force those. For the pur­poses of this post, we’re in­ter­ested in a sub­set of space gov­er­nance that I will call long-term space gov­er­nance. Long-term space gov­er­nance refers to the pro­cesses of in­ter­ac­tion and de­ci­sion-mak­ing among the ac­tors in­volved in the large-scale set­tle­ment of space.

Space coloniza­tion is cur­rently not well cov­ered by ex­ist­ing gov­er­nance mechanisms. The most sig­nifi­cant treaty in in­ter­nal space law is the Outer Space Treaty, signed in 1967, which es­tab­lishes that space shall be free for ex­plo­ra­tion and use by all na­tions, but that no na­tion may claim sovereignty of outer space or any ce­les­tial body.[1]

Sub­se­quent efforts to es­tab­lish more com­pre­hen­sive rules, such as the Moon Treaty (which grants ju­ris­dic­tion over ce­les­tial bod­ies to the in­ter­na­tional com­mu­nity), have largely failed to achieve wide­spread as­sent. There­fore, we cur­rently lack a co­her­ent global frame­work for space gov­er­nance. As of now, space is a free-for-all.[2] This is par­tic­u­larly true for challenges that arise in the con­text of hu­man­ity ex­pand­ing be­yond Earth: large-scale set­tle­ments in space are cur­rently in­fea­si­ble, so much of the ex­ist­ing de­bate cen­ters on more im­me­di­ate con­cerns (e.g. re­lated to satel­lites or ex­plo­ra­tion of space).

The work I have in mind aims to re­place the cur­rent state of am­bi­guity with a co­her­ent frame­work of (long-term) space gov­er­nance that en­sures good out­comes if and when large-scale space colon­i­sa­tion be­comes fea­si­ble. In the fol­low­ing, I will ar­gue that such work is im­por­tant, tractable, and ne­glected.


The case for the im­por­tance of space gov­er­nance is straight­for­ward: it di­rectly af­fects as­tro­nom­i­cal stakes. On a cos­mic scale, Earth is a tiny point in a vast uni­verse con­tain­ing hun­dreds of billions of galax­ies. Our own galaxy, the Milky Way, already con­tains at least 100 billion planets. So, while space gov­er­nance is not fun­da­men­tally differ­ent from ex­ist­ing gov­er­nance prob­lems, it takes place on a scale never be­fore seen in hu­man his­tory.

Also, the range of pos­si­ble out­comes is huge. The right space gov­er­nance regime could en­able an out­come that is very good from (al­most) ev­ery per­spec­tive—through pos­i­tive-sum co­op­er­a­tion and com­pro­mise be­tween the rele­vant ac­tors, com­bined with the vast amount of re­sources that an in­ter­galac­tic civil­i­sa­tion can ac­cess. (Cf. Eric Drexler’s Pare­to­topia.) On the other side of the spec­trum, es­ca­lat­ing con­flicts and war­fare on a cos­mic level could cause ac­tors to in­flict uni­mag­in­able hor­rors on each other, re­sult­ing in suffer­ing on an as­tro­nom­i­cal scale.

That said, one could ob­ject that any­thing we can do now will be over­turned in the fu­ture, ren­der­ing our efforts ir­rele­vant. In par­tic­u­lar, one might ex­pect trans­for­ma­tive AI to hap­pen rel­a­tively soon (which may be the trig­ger for large-scale space colon­i­sa­tion), and pow­er­ful fu­ture AI sys­tems may not be bound by laws (or other as­pects of gov­er­nance) in the same way as hu­mans. This is why we should (so the ar­gu­ment goes) in­stead fo­cus on shap­ing the im­pacts of trans­for­ma­tive AI.

While this is a pos­si­bil­ity, I think it is at least plau­si­ble that laws and gov­er­nance paradigms are long-last­ing, or that the fi­nal out­come is path-de­pen­dent on es­tab­lish­ing good space gov­er­nance early on. Depend­ing on what the tran­si­tion to trans­for­ma­tive AI looks like, there is at least a sig­nifi­cant chance that ex­ist­ing laws or in­sti­tu­tions re­main rele­vant, or set a di­rect prece­dent. At first glance, this seems to ap­ply to grad­ual, dis­tributed tran­si­tions in par­tic­u­lar.[3] (Also, a) similar ob­jec­tions ap­ply to many at­tempts to im­prove the long-term fu­ture, and b) I do not make the strong claim that space gov­er­nance is more im­por­tant than di­rectly shap­ing the im­pacts of trans­for­ma­tive AI.)

A re­lated ob­jec­tion is that it may not be ur­gent to work on space gov­er­nance now, as we can del­e­gate the ques­tion of how to gov­ern space to (hope­fully) more ca­pa­ble and more in­formed suc­ces­sors (whether ar­tifi­cial or hu­man). This is es­pe­cially true if (large-scale) space colon­i­sa­tion is un­likely to hap­pen soon – i.e., if our space­far­ing ac­tivi­ties will, for the fore­see­able fu­ture, be limited to ex­plo­ra­tion.

But I would, again, ar­gue that es­tab­lish­ing good space gov­er­nance is plau­si­bly time-sen­si­tive. It seems at least pos­si­ble that our civil­i­sa­tion will start set­tling other planets in the not-too-dis­tant fu­ture – say, within the next cen­turies.

One con­crete rea­son why space gov­er­nance is time-sen­si­tive is that en­forc­ing pos­i­tive-sum agree­ments may be­come more difficult or im­pos­si­ble once civil­i­sa­tion has already spread into space, due to mas­sive cos­mic dis­tances. For in­stance, it may be the case that agree­ments can only be en­forced if the re­quired surveillance and en­force­ment mechanisms are “built into” the colon­i­sa­tion pro­cess and form an in­te­gral part of each colony from the out­set (so that com­mu­ni­ca­tion or in­ter­ac­tion over vast dis­tances is not nec­es­sary). The best op­tion, there­fore, would be to set­tle on a fair gov­er­nance paradigm be­fore space colon­i­sa­tion be­gins.

I would also note that the cur­rent state of am­bi­guity seems bad for many rea­sons. It could in­duce race dy­nam­ics, where ac­tors rush to claim re­sources with­out re­gard for risks, and is likely to re­sult in se­ri­ous con­flict. There­fore, im­prov­ing space gov­er­nance may also help re­duce the risk of great power war.


It seems fairly clear to us what lev­ers we can pull on to im­prove space gov­er­nance. We can do re­search on what gov­er­nance frame­work we want, and then shape ex­pert dis­course, or lobby for the right laws and in­ter­na­tional con­ven­tions. (I will elab­o­rate on this be­low.) Also, work on space gov­er­nance does not hinge on par­tic­u­lar as­sump­tions about the fu­ture, other than large-scale space colon­i­sa­tion be­ing suffi­ciently plau­si­ble.

I do not yet have a good sense of how tractable it is to im­ple­ment space gov­er­nance im­prove­ments. One po­ten­tial prob­lem is that space gov­er­nance is mainly shaped by large space-far­ing na­tions (US, China, Rus­sia, EU, In­dia). Since they of­ten seem to have com­pet­ing in­ter­ests in this do­main, an agree­ment might be very hard to reach, es­pe­cially if re­la­tion­ships are strained for rea­sons un­re­lated to space gov­er­nance. (Per­haps this is why there has not been much progress for a long time.)

For­tu­nately, when it comes to long-term space gov­er­nance, we may be able to pull the rope side­ways since com­pet­ing state in­ter­ests mostly re­late to short-term is­sues. In ad­di­tion, non­state in­ter­ests (mainly civil so­ciety and com­mer­cial ac­tors) in space seem quite small at this point, which makes it more plau­si­ble that tar­geted work could have an out­sized im­pact. It is eas­ier to have a sig­nifi­cant in­fluence while there aren’t yet strong vested in­ter­ests. (The flip­side of this ar­gu­ment is that there may not be much in­ter­est in dis­cussing long-term is­sues in the var­i­ous gov­ern­ments.)


There is some work on space law and space gov­er­nance (e.g. by the Com­mit­tee on the Peace­ful Uses of Outer Space), but our im­pres­sion is that this is mostly fo­cused on short-term is­sues. I am not aware of any ac­tors do­ing sys­tem­atic work or ad­vo­cacy re­gard­ing the long-term con­cerns out­lined in this post, al­though the need for a co­her­ent space gov­er­nance frame­work has been pointed out (e.g. 1, 2).

What is good space gov­er­nance?

So far, I brack­eted the ques­tion of what kind of space gov­er­nance regime I would like to see. This is be­cause the long-term effects of gov­er­nance regimes are hard to pre­dict, and fur­ther re­search is needed to be con­fi­dent in any par­tic­u­lar idea. Even ab­stract­ing away from prac­ti­cal and poli­ti­cal con­straints, it’s not clear what the ideal gov­er­nance paradigm would be. Nev­er­the­less, I would like to offer some pre­limi­nary thoughts.

A top pri­or­ity is to re­duce the risk of ma­jor con­flict. Space war­fare is not only a mas­sive waste of re­sources, but might also re­sult in moral atroc­i­ties, similar to past wars—but on an as­tro­nom­i­cal scale. In other words, it is a po­ten­tial s-risk. It is clear that we would much pre­fer a gov­er­nance regime that pro­motes pos­i­tive-sum co­op­er­a­tion and com­pro­mise.

More speci­fi­cally, we would like to en­sure cos­mic rule of law. This en­tails an ad­e­quate level of ex­ter­nal con­trol—through en­force­ment of cer­tain laws, norms, and agree­ments—over what ac­tors in space are (and are not) al­lowed to do. It should be pos­si­ble to stop bad ac­tors from caus­ing a lot of harm: we should pre­vent cos­mic an­ar­chy.

This is challeng­ing be­cause vast dis­tances in space will likely be an ob­sta­cle to effec­tive en­force­ment. Space is, in a nut­shell, an end­less desert with oases that are ex­tremely far apart from each other. The clos­est star to Earth is 4.3 light years away, re­sult­ing in a round-trip la­tency of 8.6 years even if light-speed com­mu­ni­ca­tion and trans­port are pos­si­ble. The clos­est galaxy is ap­prox­i­mately 2.5 mil­lion light years away, ren­der­ing con­ven­tional en­force­ment im­pos­si­ble.

I think it is also im­por­tant to avoid a laissez-faire space gov­er­nance paradigm, where each ac­tor owns a cer­tain frac­tion of space and can do what they want within their ter­ri­tory. Such prop­erty rights would ar­guably work for com­mer­cial and self­ish in­ter­ests, and may well be the ar­range­ment that would emerge by de­fault. How­ever, this sce­nario is prob­le­matic (es­pe­cially from an s-risk per­spec­tive) be­cause it would be im­pos­si­ble to stop moral catas­tro­phes.

What would be a good space gov­er­nance paradigm? Here are some first-pass ideas (as­sum­ing a suffi­cient de­gree of en­force­abil­ity):

  • Prop­erty rights could be com­bined with a veto mechanism to stop uses of space that are con­sid­ered mal­i­cious by suffi­ciently many ac­tors.

  • Rather than al­lo­cat­ing ter­ri­tory, we could only al­low rights to use cer­tain re­sources in cer­tain ways. This could re­quire au­tho­ri­sa­tion by a cen­tral au­thor­ity (similar to e.g. plan­ning per­mis­sion for new de­vel­op­ments).

  • Rather than di­vid­ing space up, it could be con­sid­ered a com­mons with shared (demo­cratic) con­trol by many stake­hold­ers.

In any case, it would ar­guably be good to set­tle on a clear paradigm be­fore large-scale space colon­i­sa­tion be­comes fea­si­ble. Ideally, hu­man­ity would first set­tle on a fair com­pro­mise be­tween ev­ery­one’s val­ues and in­ter­ests, and then colon­ise space ac­cord­ing to this com­pro­mise. (But this would re­quire a mas­sive im­prove­ment in the (gen­eral) gov­er­nance situ­a­tion of our civil­i­sa­tion, which is ar­guably not re­al­is­tic in the fore­see­able fu­ture.)

How to work on this?

As there hasn’t been much work on long-term space gov­er­nance, a top pri­or­ity is fur­ther re­search. I sug­gest the fol­low­ing di­rec­tions:

  • Com­piling an overview of pos­si­ble gov­er­nance paradigms and how de­sir­able they are, both from an s-risk per­spec­tive and other per­spec­tives.

  • Con­sider how re­al­is­tic each solu­tion is, given the in­ter­ests of differ­ent stake­hold­ers.

  • Con­nect­ing these long-term con­cerns with the ex­ist­ing dis­course on space law, which is largely fo­cused on more con­crete, im­me­di­ately rele­vant is­sues.

  • Ex­plor­ing whether and how en­force­ment is pos­si­ble in a tech­nolog­i­cally ad­vanced, in­ter­galac­tic civil­i­sa­tion.

In ad­di­tion, I recom­mend that more peo­ple build up ex­per­tise or even con­sider a ca­reer in space gov­er­nance. Poli­ti­cal sci­ence, in­ter­na­tional re­la­tions, and law with a fo­cus on space gov­er­nance are par­tic­u­larly rele­vant. I have not con­sid­ered ca­reer op­tions in de­tail, but plau­si­ble paths in­clude aca­demic ca­reers in these fields, ca­reers in rele­vant gov­ern­ments, think-tanks, or in­ter­na­tional or­gani­sa­tions, as well as ca­reers at NewS­pace com­pa­nies. I don’t think many peo­ple in the com­mu­nity should pur­sue this path, but it could be a very promis­ing op­tion for those who are pas­sion­ate about the topic.


Ste­fan Torges con­tributed sig­nifi­cantly to this post through ex­ten­sive com­ments, in­puts and sug­ges­tions.

I’d also like to thank Lukas Gloor and Jesse Clif­ton for com­ments on an ear­lier draft of this text.

  1. Fol­low­ing the above defi­ni­tion, space law is an im­por­tant as­pect of space gov­er­nance, but the lat­ter is broader in scope. ↩︎

  2. The Com­mer­cial Space Launch Com­pet­i­tive­ness Act of 2015 ex­plic­itly al­lows US cit­i­zens and in­dus­tries to “en­gage in the com­mer­cial ex­plo­ra­tion and ex­ploita­tion of space re­sources”. It is de­bated whether this con­sti­tutes a claim of sovereignty (in vi­o­la­tion of the Outer Space Treaty). ↩︎

  3. An in-depth dis­cus­sion of the plau­si­bil­ity of differ­ent AI sce­nar­ios and how they af­fect space coloniza­tion is be­yond the scope of this text. ↩︎