By the time there is an interstellar civilization, humanity would almost certainly already have caused a superintelligence to be created, which would have plenty of time to both achieve technological maturity and superintelligent governance over the physical substrate. This makes a lot of the framing in the post moot.
So any physics-related dangers originating from within the civilization shouldn't be possible to enact, since the colonized areas of the physical world are under superintelligent control. And any external dangers likely can't have a meaningful technological advantage, since all technologically mature civilizations have a similar level of knowledge about the physical world and ability to produce any relevant physical technology. The external dangers could only have some advantage in the area of the universe already under their control.
Also, distributed backups make any local physical damage somewhat ephemeral, as long as the rest of the universe under the same governance (expanding with colonization) has enough matter to work with, to restore anything of value that got its original physical substrate taken away.
By the time there is an interstellar civilization, humanity would almost certainly already have caused a superintelligence to be created
This is definitely correct, but I think this scenario hops over the decision-point that I'm referring to (or, the point at which the governance system I describe becomes necessary). It isn't the time of the creation of interstellar civilisation that I care about, but the point at which interstellar civilisation becomes inevitable. This may be the first interstellar colonisation mission, which begins a self-propagating expansion of human civilisation throughout the galaxy. If that propagation begins without an extremely forward-thinking governance system embedded within it (seemingly superintelligent), then the eventual creation of a galactic x-risk seems inevitable.
I think it's very plausible that superintelligence comes after the first interstellar mission. One reason is that transformative AI, initiating explosive technological growth, would presumably come before superintelligence. So we would probably gain the ability to send an interstellar colonisation mission at that point before the development of superintelligence.
To be fair, we would almost certainly be able to catch up with an interstellar colonisation mission post-superintelligence, which might follow it quite quickly. Though, I'm not sure if this is the best way to go about ensuring long-term existential security, as they may be resistant to being governed by an AI. So I would prefer to play it safe and avoid interstellar travel completely until the problems I outlined are solved - that's more of a personal preference than a strong opinion though.
And any external dangers likely can't have a meaningful technological advantage, since all technologically mature civilizations have a similar level of knowledge about the physical world and ability to produce any relevant physical technology.
I'm not sure if I understand this. Are you saying that if an external civilisation (i.e., alien civilisation) initiates a galactic x-risk, then another civilisation governed by a superintelligent AI would be able to protect itself because it would have similarly mature technology? I'm not sure if there is anything we could do to defend ourselves if an alien civilization initiated vacuum decay.
Also, distributed backups make any local physical damage somewhat ephemeral, as long as the rest of the universe under the same governance (expanding with colonization) has enough matter to work with, to restore anything of value that got its original physical substrate taken away.
Yeah, definitely. All of the galactic x-risks I listed though are either self-propagating or infinite in range, except maybe societal collapse or loss of value scenarios.
I would add that there are a series of planetary system-wide risks that appear only for civilizations traveling within their solar systems but do not affect other solar systems. These include artificial giant planet explosions via initiating nuclear fusion in their helium and lithium deposits, destabilization of the Oort cloud, and the use of asteroids as weapons.
More generally speaking, any spacecraft is a potential weapon, and the higher its speed, the more dangerous it becomes. Near light-speed starships are perfect weapons. Even a small piece of matter traveling at very high velocities (not necessarily light speed, but above 100 km per second, as I recall) will induce large nuclear reactions at the impact site. Such nuclear reactions may not produce much additional energy but will cause significant radioactive contamination. They could destroy planets and any large structures.
Additionally, space colonization will likely develop alongside weapon miniaturization, which could ultimately result in space-based grey goo with some level of intelligence. Stanisław Lem's last novel, "Fiasco," seems to address this concept.
I think people are very wrong about space colonisation increasing long-term existential security. Once we expand to other star systems, we may begin a self-propagating expansion of human civilisation throughout the galaxy. However, there are existential risks potentially capable of destroying a galactic civilisation, like self-replicating machines, strange matter, and vacuum decay. Without an extremely widespread and effective governance system, the eventual creation of a galaxy-ending x-risk seems almost inevitable due to cumulative chances of initiation over time across numerous (potentially hundreds of thousands of) independent actors. The requirements for a governance system to prevent galactic x-risks are extremely demanding, and it needs to be in place before interstellar colonisation is initiated.
Introduction
I recently came across a series of posts from nearly a decade ago, starting with a post by George Dvorsky in io9 called “12 Ways Humanity Could Destroy the Entire Solar System”. It’s a fun post discussing stellar engineering disasters, the potential dangers of warp drives and wormholes, and the delicacy of orbital dynamics.
Anders Sandberg responded to the post on his blog and assessed whether these solar system disasters represented a potential Great Filter to explain the Fermi Paradox, which they did not[1]. However, x-risks to solar system-wide civilisations were certainly possible.
Charlie Stross then made a post where he suggested that some of these x-risks could destroy a galactic civilisation too, most notably griefers (von Neumann probes). The fact that it only takes one colony among many to create griefers means that the dispersion and huge population of galactic civilisations[2] may actually be a disadvantage in x-risk mitigation.
In addition to getting through this current period of high x-risk, we should aim to create a civilisation that is able to withstand x-risks for as long as possible so that as much of the value[3]of the universe can be attained as possible. X-risks that would destroy a spacefaring civilisation are important considerations for forward-thinking planning related to long-term resilience. Our current activities, like developing AGI and expanding into space may have a large foundational impact on the long-term trajectory of human civilisation.
So I've investigated x-risks with the capacity to destroy a galactic civilisation[4] ("galactic x-risks"), defined here as an event capable of destroying the long-term potential of an arbitrarily large spacefaring civilisation[5]. A galactic x-risk would be a huge moral catastrophe. First, that's a lot of death. Second, once a civilisation has overcome the barriers required to become a galactic civilisation, their long-term potential in expectation is probably much higher than a civilisation on one planet (i.e., a pre-Precipice or pre-Great Filter civilisation). So, the loss of future value may be much greater[6].
Existential risks to a Galactic Civilisation
I'll start by ruling out the risks that I believe are limited in scope to a one planet civilisation, and then to a small spacefaring civilisation. Don't worry too much about the categories/subheadings here though. These threats could be in multiple categories depending on their severity, and, if artificial, their design and deployment strategy[7]. The threats in the "Galactic Existential Risks" section will be the most relevant to the long-term resilience of a galactic civilisation.
I have not classified existential risks into artificial and natural because we’re dealing with arbitrarily advanced technology. So any natural event that we know can occur, could conceivably be induced by arbitrarily advanced technology, or at least the probability of its occurrence could be increased artificially.
Threats Limited to a One Planet Civilisation
The threats we will escape by becoming spacefaring are those that drastically change a planet’s atmospheric conditions without spreading to space. Included in this group are volcanic eruptions[8], geoengineering disasters[9], atmospheric positive feedback loops[10], the release of highly toxic molecules/chemical warfare[11], nuclear war[12], and an asteroid or comet impact[13]. By spreading outside of one’s own atmosphere, other settlements can survive and hopefully safeguard the long-term potential of human civilisation[14].
Threats to a small Spacefaring Civilisation
Some existential risks can propagate through space and potentially destroy civilisations across multiple star systems. However, threats that are not self-propagating or are limited in range can eventually be overcome by spreading throughout the galaxy. These threats, therefore, only pose existential risks during a relatively brief phase in the expansion of human civilisation in space.
The longest list of threats in this transitionary phase come from stars. Firstly, a civilisation's host star may be the source of x-risks like superflares[15], stellar mass loss events[16], increasing stellar luminosity or volume, or stellar engineering disasters[17]. Alternatively, localised existential risks could come from natural energetic events of other stars[18], like supernova explosions, magnetar flares, pulsar beams, kilonova winds[19], and M-dwarf megaflares. These energetic stellar events would affect multiple star systems, with a range of impacts depending on the type of event and its distance.
There are also all sorts of cosmic phenomena that could be encountered by a civilisation occupying one or multiple solar systems (threat-dependent) that would likely destroy them. These include wormholes[20], primordial black holes[21], cosmic strings[22], domain walls[23], interstellar clouds[24], and supernovae remnants[25].
Alternatively, any artificial or natural[26] alterations to the orbits of planets could conceivably destroy a spacefaring civilisation around one star by altering UV radiation influx or inadvertently causing planet collisions[27].
Giant artificial structures in space like Dyson spheres, orbital rings, sunshades, Shkadov thrusters, or space particle accelerators could conceivably end a spacefaring civilisation too. Their immense energy demands, gravitational influence, or stored kinetic potential makes them capable of causing catastrophic failure modes, like stellar destabilisation, orbital cascade, or directed energy misfires[28]. Additionally, interstellar weapons like high powered lasers and particle beams may be used to annihilate a civilisation occupying a solar system from very far away[29].
The most devastating localised existential risks are galactic core explosions and quasars[30]. A quasar could generate a galactic superwind from the center of a galaxy, stripping away atmospheres as it expands, sterilising large regions of the inner galaxy.
Galactic Existential Risks
This is a list of threats that could conceivably destroy the long-term potential of a whole galactic civilisation, or plausibly even a civilisation occupying multiple galaxies. Preventing any of these from occurring[31] is essential to the long-term resilience of a galactic civilisation.
all it takes is one civilization of alien ass-hat griefers who send out just one von Neumann Probe programmed to replicate, build N-D lasers, and zap any planet showing signs of technological civilization, and the result is a galaxy sterile of interplanetary civilizations until the end of the stelliferous era (at which point, stars able to power an N-D laser will presumably become rare).
I think the key here is self-replication. If technologically feasible, self-replicating interstellar lasers are much worse than interstellar lasers. Presumably, to annihilate a galactic civilisation, the creator of the von Neumann probe would also be eliminating themself. But I think that releasing von Neumann probes that kill most civilisations and prevent others from emerging is sufficient to destroy most of the long-term potential of a galactic civilisation[33]. In the words of Charlie Stross, “it only has to happen once, and it f***s everybody".
I'm also including nanotechnology in the self-replicating machines section. There are definitely motivations to decrease the size of a von Neumann probe. If the probes are adaptive, then nanotechnology might be an effective strategy at times. So the same overview above applies to nanotechnology.
Biological machines (i.e., engineered pandemics) are also self-replicating machines, but are probably not an effective strategy for deployment in space. One could argue that weaponised bioengineered pandemics could be a threat to a spacefaring civilisation if they were adaptive[34], engineered for maximum spread and lethality, and deployed artificially across the galaxy[35].
Strange matter
Strange matter has been theorised to exist as a more stable form of matter, and it may be able to turn normal matter into strange matter on contact. Strange matter is predicted to form when atoms are compressed beyond a critical density, dissociating the protons and neutrons in quarks, creating quark matter and strange matter. These conditions might be reached during a collision between two neutron stars, and strangelets (small droplets of strange matter) may be released. Alternatively, extremely advanced technology may be used to create strange matter in the long-term future.
If a strangelet hit the Earth, we'd all be converted into strange matter, and strangelets would be created and dispersed. Eventually this process of conversion and spread could destroy a galactic civilisation.
Strange matter is probably a solvable risk for an extremely advanced civilisation. With huge amounts of energy, they may be able to create gravitational waves to act as a repulsor beam. But still, a huge amount of long-term value in the universe may be lost if the strange matter is not contained rapidly.
This would result in an expanding bubble of total destruction that would sweep through the galaxy and beyond at the speed of light, tearing all matter apart as it proceeds.
Very speculative grand unified theories in the 1970s (e.g. SU(5)[37] and SU(10)) implied that subatomic particles can decay[38], meaning all matter is ultimately unstable. For example, the proton may decay via an interaction with a magnetic monopole[39] or via pathways involving virtual black holes and hawking radiation[40]. Additionally, neutrons may decay[41] via pathways like neutron-antineutron oscillationsor by leaking into other dimensions[42]. However, these processes mainly act locally, so they are unlikely to be galaxy-ending scenarios unless they were self-propagating reactions[43] or another fundamental process altered physics (either artificially or naturally over time) to allow the decay to occur across the universe[44].
Time travel
In general relativity, if it's possible to create spacetime warping structures like cosmic strings and wormholes, then time travel back in time is not definitely impossible. There are some rebuttals like chronology protection from quantum physics and the chronology protection conjecture. But, generally, it seems impossible to know for sure whether time travel is actually possible until a unifying theory to join quantum gravity with quantum mechanics and general relativity has been developed[45]. If time travel is possible, then (depending on your personal favourite theory of time travel) a galactic civilisation might be destroyed by itself before it is created.
Fundamental Physics Alterations
There are so many different fundamental constants and properties that are highly precise and necessary to the existence of life. With extremely developed physics and arbitrarily advanced technology, if any of them are possible to alter at a large scale, then life in the universe is at risk. Additionally, it has been suggested that fundamental constants may drift over time in accordance with the age of the universe. The fundamental constants include the gravitational constant[46], strong coupling constant[47], Planck’s constant[48], the cosmological constant[49], and the fine structure constant[50]. Other stuff breaking down that would be very bad include color confinement[51], quark-leption unification[52], and the speed of light.
Interactions with Other Universes
If there are other universes or our universe is a sub-universe[53], then everything could end quite abruptly.
According to brane theory, other universes exist in non-visible dimensions, but are able to collide with our universe (brane collision)[54]. One of these collisions might have ignited the Big Bang. If a Brane collided with our universe again, it could initiate another Big Bang-like event and reset our universe.
Alternatively, our universe could actually exist in the black hole of another universe. Black holes contain spacetime singularities, which break down everything we know about space and time, so space and time may switch roles. From the outside, a black hole is infinitely large in time (i.e., it may continue existing almost indefinitely), but from the inside it is infinitely large in space. That means a black hole could contain a whole universe... our universe. This might not actually change much about our understanding of the universe. However, the collapse of a black hole universe into a singularity would probably end our universe in a similar way to the Big Crunch[55]. Alternatively, our host black hole might just evaporate slowly by releasing hawking radiation, this would normally take 1067 years, far longer than the current age of our universe (1.38 x 1010).
Finally, (I say in a slow rhythmic voice) if the universe is a simulation and the simulation was turned off, a civilisation at any scale in our universe would be destroyed.
Societal Collapse or Loss of Value
Galactic civilisations may fizzle out due to societal collapse or loss of their moral value. Societal collapse may occur via predator prey dynamics[56], galactic tyranny, weakness by homogeneity[57], or an interstellar coordination breakdown. A loss of moral value might come about by an outcompeting of sentient beings by non-sentient posthumans or artificial intelligence, the creation of s-risks, or a gradual drift of values.
Artificial Superintelligence
Artificial superintelligence can use all of the above to kill a galactic civilisation.
I think the creation of superintelligent AI may always be an x-risk, even if alignment has been solved previously somewhere else in the universe.
While war has typically favoured the defender, destructive capabilities have always been more powerful than defensive or constructive capabilities. While countries have been created over thousands of years, any country on Earth could be destroyed within the next hour by nuclear weapons. So I would assume that, at the limits of technology, destructive capabilities will eventually be so great as to defeat any defences[58], vacuum decay being the prime example.
So even if an aligned superintelligence exists in a galactic civilisation. Then assuming new civilisations are able to emerge independently of it, the creation of an evil superintelligence somewhere else is probably still an x-risk as its destructive nature would make it innately more powerful than the benevolent AI. Especially if the evil AI has no interest in self-preservation (e.g., it was created to eliminate astronomical suffering), it could initiate any number of x-risks that there is no defence against, like vacuum decay, N-D lasers, von Neumann probes equipped with N-D lasers, strange matter, or likely many more things that I couldn't even conceive of.
One potential solution to prevent an evil superintelligence from emerging is to have an aligned superintelligence that is able to observe all activities across a galaxy (likely requiring independently acting but coordinated superintelligences to mitigate communication delays). This way, the emergence of a new superintelligence could be prevented, along with all the other galactic x-risks on the list. More on this in the last section as there are obvious downsides to address.
Conflict with alien intelligence
There are essentially infinite ways an alien civilisation with arbitrarily advanced technology might choose to eliminate a galactic civilisation[59]. Just pick any item on the galactic existential risks list and insert before it "aliens decide to use...", or pick anything in the previous section (threats to a small spacefaring civilization) and insert it into the blank: "aliens decide to use _______ millions of times".
I think aliens are more dangerous than superintelligence. Firstly, because aliens could create superintelligence. Secondly, we may have no control over their creation or what they might do. They could just initiate vacuum decay right now and we wouldn't even see it coming as it approaches at light speed. More on aliens in the very uplifting section entitled: "If aliens exist, there is no long-term future".
Unknowns
Assuming that humanity right now is nowhere near the limits of technological and scientific knowledge, a post-ASI civilisation could expand the number of galactic x-risks drastically. If you asked a 21st-century PhD student (me) and a superintelligent AI to list all possible galaxy-ending existential risks, what percentage of the AI's total list would the PhD student be able to name? Asking us to list galaxy-ending x-risks today might be like asking an ancient Roman to predict how artificial intelligence could collapse 21st-century civilisation - the concepts simply don’t exist yet. I wouldn't be so bold as to claim I could name more than 50% of the galactic x-risks.
Figuring out how many unknown unkowns there are is really hard. The sample size of galactic x-risks isn't really big enough to say anything confidently about a rate of discovery, especially since I don't know which are actually real. So I've listed 100 cosmic threats[60] (anything arising from space that could destroy Earth as a proxy for galactic existential risks), summed up how many were discovered in each decade, and plotted that on a bar graph[61]:
Aside from this just being super interesting, there are signs we may be reaching diminishing returns - it looks like we've passed the peak of a bell curve[62]. However, unexpected discovery classes remain possible, and the discovery rate may go up rapidly if there is an intelligence explosion. How high is pure speculation I think. There is so much that we know we don't know about the universe, so there must be so much that we don't know we don't know. But playing it safe, I'd assume that I've listed less than 50% of the actual number of galactic x-risks, and we should probably act as if we're certain that there are many galactic x-risks.
What is the probability that galactic x-risks I listed are actually possible?
For an x-risk on the list to actually be a galactic x-risk, it must be real and capable of destroying a galactic civilisation. Here are my best guesses for each of the galactic x-risks[63]. For each, I have given my percentage probability of the threat actually being a thing that could happen in the long-term future and of the threat being capable of ending a galactic civilisation if it was initiated. Threats resolving 100% on either get a tick.
Reassuringly, none of these have two ticks in my estimation. However, combined, I think this list represents a threat that is extremely likely to be real and capable of ending a galactic civilisation.
What is the probability that an x-risk will occur?
What are the factors?
There are multiple factors that simultaneously increase and decrease x-risk as a civilisation expands in space:
Dispersion: As a colony expands into space, it becomes more dispersed, which reduces the probability that a single x-risk will destroy everyone. However, dispersion across interstellar space makes effective governance increasingly challenging due to the huge communication time lags[66].
Population size: Similar to dispersion, the more sentient beings that exist, the higher the probability of survival. But the higher the probability that one of those individuals or civilisations will cause an x-risk - it only takes one to initiate a galactic x-risk.
Resource availability: As we expand into space, we are able to take advantage of more and more energy to simultaneously create much greater weapons and defences against x-risks.
I think that, historically, the negative sides of dispersion and population size have not been appreciated fully when considering existential security benefits from space expansion.
Cumulative Chances
The initiation of a galactic x-risk could be motivated by a desire to end ongoing astronomical suffering across a galaxy, which I think is the most likely scenario. But generally, I think it doesn't really matter what motivations there are because even very small probabilities of an event occurring can become extremely large given enough time. Assuming that galactic x-risks are possible (i.e., they exist and can destroy a galactic civilisation), cumulative chances over time and space make the probability of a galactic x-risk occurring nearly 100%.
If an event has an n > 0 probability of occurring over a period of time P(t), then the probability of that event having occurred over increasingly long time periods will eventually approach 100% (the green line in the graph). Additionally, the probability of an event occurring increases proportionally to the number of actors who are capable of initiating that event (as indicated by the other two lines in the graph). A galactic civilisation with hundreds of millions of independent star systems will be doomed relatively quickly by a correlated x-risk. This graph is adapted from a talk that Toby Ord gave at EAG London 2025 on forecasting over long time periods:
So, for a galactic civilisation[67], even if the probability of a star system inducing a galactic x-risk is very low, the probability that a galactic x-risk will eventually be triggered is effectively 100%. This poses a potentially enormous threat to the total value that human civilisation could gain from the universe.
If aliens exist, there is no long-term future
I hinted at this in the 'conflict with alien intelligence' section. I have suggested that galactic x-risks are real and could destroy a spacefaring civilisation of arbitrarily large size. I then argued that the probability of one of those risks occurring is extremely high. So if aliens exist at the same time as us (especially under a grabby aliens scenario), then the probability that they will initiate an x-risk that would affect us is also very high, whether intentionally or not. There might already be vacuum decay bubbles headed our way, so human civilisation would end even if we did everything right. In addition to concerns I'm raising about the resilience of our future galactic civilisation, I think aliens have big implications for discussions around big picture cause prioritisation[68].
If alien-induced galactic x-risks have a non-zero probability of occurring, it's plausible that interstellar civilisations facing similar existential risks might develop coordination mechanisms or mutual deterrence strategies over time. And we shouldn’t rule out the possibility that sufficiently advanced alien civilisations could be benevolent or even value-aligned in ways that reduce risk rather than exacerbate it. But the risk is very high if they will emerge and spread independently of us - the cumulative probabilities graph above is multiplied and out of our control.
The Way Forward
From hereon I will assume we are the only intelligent life in the universe at the moment. What's our best path?
The governance systems of a galactic civilisation would have to be amazing to prevent some of these galactic x-risks. This problem might not be solved or be given sufficient attention within the next few decades. However, once an interstellar mission is sent (Metaculus predicts 2116 (25%), 2248 (median), >2500 (75%)[69]), a self-perpetuating expansion of humanity throughout the cosmos is plausibly initiated. Taking time for a long reflection before initiating such an expansion seems important for increasing the chances of exporting a viable galactic governance system - one capable of sustaining a flourishing civilisation that can endure and spread across the galaxy for billions of years.
One solution is to abandon galactic colonisation and only expand in the digital space, using the real world exclusively for gathering resources[70]. But if we choose galactic colonisation, then a governance system would need to meet the following requirements in order to prevent galactic x-risks:
The governance system should be able to spread ahead of (or at least with) human civilisation throughout the affectable universe.
The governance system should be impossible to overthrow, manipulate, sidestep, or avoid.
All potential routes to the creation of a galactic x-risk or astronomical suffering should be known by the governance system.
Any activities pertaining to potential routes to galactic x-risks or astronomical suffering should be observable by the governance system.
The governance system should hold or have access to the powers to prevent any actor from initiating a galactic x-risk or astronomical suffering.
The governance system should be able to identify emerging alien civilisations and integrate itself into them or collaborate with them to continue meeting the above requirements.
It's worth noting that conditions 3 to 5 basically describe God - all knowing and all powerful. So it might be necessary to create a God-like being, or at least a superintelligence, prior to interstellar colonisation. The amount of foresight required to meet these conditions is inconceivable without a superintelligent system that is able to adapt to emerging hazards (or fundamentally unknowable future scenarios like other universes interacting with ours). It seems impossibly hard to predict what any one of hundreds of millions of star systems developing across a galaxy for billions of years might do. The good news is that we are in a very powerful position right now, especially if we are alone in the universe. We could prevent any star system in the future from gaining more power than their overarching governance system. This advantage is lost if humanity spreads to other star systems soon and allows them to culturally drift or remove themselves from a centralised governance system.
Another point to mention from the requirements is that I've included the prevention of astronomical suffering. Inadvertently locking human civilisation in a state of astronomical suffering with no way out is literally the worst possible scenario. There are situations where the initiation of a galactic x-risk may be morally preferable. So, to avoid galactic x-risks, astronomical suffering should necessarily be prevented. There may even be an option to include a back door to allow galactic x-risks if the governance system fails to prevent astronomical suffering.
This 'governance system' I'm describing has very authoritarian vibes. Of course, pairing this with a long reflection may allow us to solve issues around freedom and governance associated with an AI overlord, or find alternative solutions. Or maybe we could become so wise that the probability of us or any of our descendants initiating a galactic x-risk is 0%.
One could argue that superintelligent AI will solve these problems, so we don't need to bother now - it would be best to focus on AI alignment. However, interstellar missions may come before aligned superintelligence. In fact, humanity would likely be able to launch an operation to colonise the entire reachable universe quite easily if we wanted to. So an interstellar mission may have to be actively prevented until we can be certain that its accompanying governance structure would either prevent further cosmic propagation or future galactic x-risks.
However, the point of no return might not be "an interstellar mission is launched". If we launch an interstellar mission in the next couple of decades, it would likely not reach anywhere near light speed and may take decades to reach its destination. So future missions with more advanced technology and faster spacecraft could likely catch up to it, even if they leave many years later. Additionally, the first interstellar settlers probably wouldn't be thinking about moving on to the next star system as soon as possible. Or, at least they probably wouldn't spread as quickly as they possibly could. So, a mission to catch up with them post-transformative AI and invite them to participate in our "governance system" would likely be successful[71]. This pathway has bad vibes though, and other star systems could be defence-dominant in the short-term[72]. In any case, I think a lot more work on investigating current plans and long-term scenarios for interstellar travel now is well justified[73].
Acknowledgements: Thanks to Jess Riedel and Toby Ord for comments on the draft. I made a lot of changes since their comments and all mistakes remain my own.
These solar system x-risks were unlikely to be future great filters as they mainly required extremely advanced technology (e.g. stellar engineering), far more advanced than interstellar travel. So, spacefaring civilisations would most likely be interstellar before any of these large-scale sci-fi disasters occurred and so they don’t represent a reliable Great Filter.
It's a combination of both these factors. A huge population with a lot of different civilisations that will be independent actors because of the huge space between them. It means that they don't act as one civilisation that could do a bad thing, but potentially millions of civilisations that could do a bad thing. In the latter scenario, the bad thing has a much higher probability of occurring... it only takes one.
I don't think there's a good word for a civilisation spanning a huge amount of space. Galactic x-risk has a good ring to it, but most of the threats I'll discuss don't care about how big your civilisation is. "Universal x-risk" isn't as clear and "cosmic x-risk" sounds too much like cosmic threats.
This definition is, of course, based on the definition of x-risk used by Toby Ord in The Precipice.
Large magnitude eruptions may dramatically cool our climate through the release of ash and sulfur, or warm the planet from significant releases of CO2. But other impacts include:
ozone destruction, reduction in rainfall, further biodiversity collapse/environmental damage (already in decline), widespread blockages of maritime trade, destruction of communication, and technology, financial losses, may lead to mass shortages of food, water and energy resources
All of these impacts combined are certainly a global catastrophic risk, and may be sufficient to cause the collapse of a global civilisation and the loss of its long-term potential.
No volcano in the Solar System is powerful enough to affect the habitability of multiple planets. However, a related but highly speculative scenario might be the detonation of a gas giant planet by the nuclear fusion of deuterium (e.g. of Jupiter). This may be initiated by a black hole impacting the planet or by the deployment of an initiator bomb[74]. The ignition of a gas giant planet may create an explosion so great that a whole solar system would be affected by it. But I'm hard-pressed to see how even this stretch of the imagination could be an existential threat to a whole spacefaring civilisation.
A geoengineering disaster may be initiated with a desire to cool the atmosphere of a planet, potentially by use of aerosols. It’s clear to see how this could go wrong. The best mitigation for this is to not do it, or make sure that if geoengineering is necessary, it is executed extremely responsibly.
The equivalent of this for a Solar System civilisation and a galactic civilisation are a stellar engineering disaster and a galactic core engineering disaster, respectively. Or maybe a planetary core engineering disaster is the more accurate planetary equivalent. Maybe. Moving on.
Including the runaway greenhouse effect and runaway refrigerator effect.
An artificially induced runaway greenhouse effect could occur if large-scale activity released vast quantities of greenhouse gases into the atmosphere. As heat becomes trapped more effectively, surface temperatures would escalate, causing oceans to evaporate and the water vapor to amplify the greenhouse effect further. This positive feedback loop could render the planet uninhabitable (Venus’ fate), collapsing ecosystems, infrastructure, and ultimately ending civilization. This is not even remotely a scenario for the present-day climate crisis.
There's also the "runaway refrigerator effect", where a positive feedback loop emerges that drives temperatures down[75]. This effect has been responsible for past ice ages on Earth and Mars.
The release of large quantities of highly toxic molecules into the atmosphere (e.g., through overenthusiastic chemical warfare) may destroy the long-term potential of human civilisation by eradicating the vast majority of the population. Most space outside of Earth's atmosphere is essentially toxic to humans anyway, so this isn't a big deal for a spacefaring civilisation.
Nukes do work in space, but a spacefaring civilisation is not as vulnerable to this because firing missiles at every space station, moon, and planet to destroy everyone would be very difficult. A self-sustaining colony capable of preserving the civilisation’s long-term potential is very likely to survive. Other more powerful interstellar weapons are discussed later.
An asteroid hitting a planet has a proven capability to cause mass extinctions and a large impact could potentially destroy human civilisation.
By expanding to other planets, moons, and space stations, we prevent all of humanity from being wiped out by one impact.
Some extreme scenarios like a strong gravitational wave from a nearby and massive cataclysmic event like a black hole merger (not gonna happen[76]) creating a large scale instability in the asteroid belt and Oort cloud could result in chaos that might destroy a spacefaring civilisation. But in those speculative scenarios, the asteroid impacts are more like an effect than the existential risk itself.
Solar flares and coronal mass ejections from stars are fairly common, and they can damage infrastructure like electrical grids and satellites.
Some have suggested that extremely powerful solar flares might also occur, so-called “superflares”. If the superflare were powerful enough, it may be capable of destroying a spacefaring civilisation by altering the surface temperatures of planets or destroying ozone layers. Humans living in space colonies may be more vulnerable to super flares if they exist outside the protection of an atmosphere. An advanced civilisation might easily mitigate the effects of superflares by predicting them in advance and using radiation shielding, or even constructing large protective structures in space.
A sudden stellar mass loss event would reduce the luminosity of the Sun. These are more common in the later stages of a star's life (and for larger stars). A mass loss event could be very destructive to a spacefaring civilisation was taken off guard. A sudden stellar mass loss event may also have other destructive consequences from the ejection of plasma and particles. However, in general, this would be predicted quite easily, is probably not a plausible scenario for the Sun, and the catastrophic consequences may be prevented with solar shields in space, stellar engineering, geoengineering, or planetary orbit alteration. A similar argument also applies to the increasing luminosity of the sun over billions of years.
The Sun may also be damaged or altered by a stellar engineering project, such as an attempt to extend the lifetime of a star or extract material, with resulting (presumably) unanticipated and destructive consequences. Additionally, an attempt to turn Jupiter into a star (e.g. to terraform the Galilean moons) may destroy Jupiter and wipe out a spacefaring civilisation. The proposed method for this was to seed Jupiter with a primordial black hole. Sounds safe.
These events have the capability to affect large regions of space. The effectors include microwave radiation, x-rays, gamma-ray bursts, cosmic radiation and particles, and neutrino showers. Space colonies may be more vulnerable to stellar explosions if they exist outside the protection of an atmosphere.
Some quantum gravity models allow for the spontaneous temporary formation of microscopic or macroscopic wormholes, which would quickly collapse, potentially into a black hole. If a wormhole were to spontaneously form near or within our Solar System and then collapse abruptly, it might produce extreme gravitational disturbances or tear spacetime locally.
dump one end in the Sun and another elsewhere (a la Stephen Baxter’s Ring), and you might drain the Sun and/or irradiate the Solar System if it is large enough.
Primordialblack holes are a theoretical type of black hole that may have formed shortly after the Big Bang. The universe was very dense then, so black holes were forming everywhere and at a range of masses. So there could still be football-sized black holes lurking about the universe. If one were to hit us at high speed, it could destroy a planet.
While physicists predict that primordial black holes would move very quickly, it's a very different story if the primordial black hole approaches slowly. In that case it could settle in a planet’s core, which it would devour from the inside. Alternatively, if primordial black holes evaporate nearby, their intense burst of radiation during final stages could cause significant damage locally.
So it seems that these primordial black holes could get up to all sorts of mischief that may destroy a civilisation spread across a solar system. Though, primordial black holes remain highly theoretical and their impacts are entirely speculative.
Cosmic strings are one-dimensional defects or cracks in spacetime that stretch for potentially millions of lightyears. They are thought to have formed from the trapping of energy into strings during early universe phase transitions. They are hypothetical but are predicted from the Big Bang and may be detectable. Simulations show that a string crossing Earth would induce "global oscillations" (like an earthquake (non-catastrophically) ringing the whole planet). Depending on the density of the strings, this could conceivably shatter Earth or the Sun. There are probably a bunch of other lethal scenarios, including but not limited to bursts of high energy particles, intense gravitational waves, and distortions of spacetime.
The early universe likely underwent multiple phase transitions, during which the fundamental fields of physics changed. Domain walls could have formed when different regions of the universe settled into different vacuum states, creating two-dimensional topological defects at the boundaries between them. But it doesn't seem that they would be stable until the present because they would be so deadly that if they were formed during the early universe, they probably would have destroyed the universe by now. But if they were stable, and we passed through one, a spacefaring civilisation would be at the mercy of whole new fundamental physics, which isn't good for biochemistry.
Interstellar clouds dominated by gas may enrich atmospheres in unusual elements, or interstellar dust clouds could generate meteor showers and destroy satellites and space stations.
Interstellar clouds would usually be a mix of gas and dust.
Supernova remnants are basically a type of interstellar (plasma) cloud.
Supernova explosions are one of the most explosive events in our universe, and they leave behind highly energetic remnants. Particularly dangerous remnants are left when pulsars go supernova, as they can contain pulsar wind nebulae, which are some of the most energetic objects in our universe (though only dangerous at short distances). If a solar system were to pass through a supernovae remnant (not gonna happen to us), then chaos would ensue for the civilisation that inhabited it.
A rogue planet entering our solar system may naturally produce these catastrophic scenarios, throwing orbits into chaos. NASA estimates that there are far more rogue planets in our galaxy than planets orbiting stars. This risk is also very conducive to extremely unlikely imaginative scenarios, like how does that change if a gas giant instead of a rocky planet enters the Solar System? What about a “rogue” black hole (or maybe even a primordial black hole)?
For this to happen accidentally, there would have to be some extreme oversight. If you have the power to move planets, you can do the maths. Also, moving planets wouldn't be the most subtle military strategy in an interplanetary conflict. So I think the natural scenario (i.e., a rogue planet) is most likely.
Their strategic value would probably make them targets in conflict, or their control could be seized by rogue actors or misaligned AI systems, leading to civilisation-wide collapse. So a spacefaring civilisation would need to incorporate very well-thought out safeguards and protections into the design of a giant structure.
Though, interstellar weapons are the kings of "use it a bunch of times and it's a galactic x-risk". I'm assuming this comes under "self-replicating machines" though. Don't worry too much about the categories.
The inflow of gas into the black hole releases a tremendous amount of energy, and a quasar is born. The power output of the quasar dwarfs that of the surrounding galaxy and expels gas from the galaxy in what has been termed a galactic superwind
This "superwind" drives all gas away from the inner galaxy. Most galaxies have already gone through a quasar phase as they were common in the early universe.
Another potential threat from the centres of galaxies are cosmic rays from galactic core explosions. Hundreds of thousands of times more damaging than supernova explosions, some theorise that cosmic rays from other galaxies have already caused mass extinctions on Earth. Explosions from the core of the Milky Way likely make the inner galaxy uninhabitable, so large explosions may have catastrophic impacts for a galactic civilisation.
Say we send out a fleet of exponentially self-replicating von Neumann probes to colonize the Galaxy. Assuming they’re programmed very, very poorly, or somebody deliberately creates an evolvable probe, they could mutate over time and transform into something quite malevolent. Eventually, our clever little space-faring devices could come back to haunt us by ripping our Solar System to shreds, or by sucking up resources and pushing valuable life out of existence.
There is an important question about offence-defence balance here. If von Neumann probes are already spread throughout a galaxy and an alien civilisation emerges, the emerging civilisations can be systematically destroyed before they become a threat.
However, if star systems are defence dominant, then it may not be a threat if one star system among a galactic civilisation releases a von Neumann probe. I am a minority viewpoint among people I've talked to about this in believing that cosmic warfare would be offence dominant. I cannot conceive of any defence against a self-replicating probe that is able to observe your defences and re-design itself accordingly with arbitrarily advanced weapons like interstellar lasers. I played this game as a kid, there is always a more powerful weapon to overcome your defence. In the plateau of technological innovation and resource availability, offence will always win. Change my mind.
A galactic civilisation would have a huge biological diversity (or even digitally sentient beings). No single pandemic would be able to infect all of them unless a superintelligent being was repeatedly designing them. In which case, superintelligence is the x-risk, and also it would have more efficient ways to kill everyone than generating plagues.
I'm definitely leaning into superintelligence for these last two points on nanotech and biological machines. Don't worry too much about the categories - the biggest galactic existential risk is a combination of multiple items on the list.
SU(5) is disproven because observational evidence indicates that the proton's half life is way longer than the theory predicts. Proton decay was a key prediction of the theory.
Other particles like photons could also decay into lighter unknown particles in speculative scenarios like Lorentz-violating theories or hidden sectors.
The gravitational constant is a fundamental physical constant that quantifies the strength of the gravitational force between objects. It appears in Newton's Law of Universal Gravitation and Einstein's theory of General Relativity. Some theories like Scalar–Tensor Theories, String Theory, Braneworld Models allow the gravitational constant to change.
This has huge implications for end-of-universe scenarios as it is closely associated with dark energy. If the total mass-energy density of the universe is high enough (i.e., if dark energy weakens), the expansion could eventually slow down and stop. If gravity overcomes dark energy, the universe would begin to contract. Eventually, the universe would shrink into a singularity - a state of infinite density and temperature, similar to the conditions at the moment of the Big Bang.
This constant governs the strength of electromagnetic interactions. Life as we know it might not be possible under even modest changes in the fine structure constant, as protein folding, DNA stability, and biochemical reactions depend on electromagnetic interactions.
In some Grand Unifying Theories, quarks and leptons are two sides of the same field. This could allow for quark-to-lepton transitions (e.g. a neutron turning into an antineutrino). Not good.
"So, bottom line, the future great filter scenario that most concerns me is one where our solar-system-bound descendants have killed most of nature, can’t yet colonize other stars, are general predators and prey of each other, and have fallen into a short-term-predatory-focus equilibrium where predators can easily see and travel to most all prey. Yes there are about a hundred billion comets way out there circling the sun, but even that seems a small enough number for predators to careful map and track all of them."
Maybe some kind of thing has ultimate moral value and is optimised for across the galaxy e.g. a simulation of digital sentience in bliss. This means that the society no longer has heterogeneity that would protect it against different types of threats. So the society may be very weak to one particular thing, like an advanced computer virus-type threat or an evil superintelligence.
I have a lot of uncertainty about this. I'd love to hear another take.. it's stuff like vacuum decay that concern me. Vacuum decay is utterly impossible to defend against, unless you could separate your star system from the rest of the universe or move into another dimension. Similar thing with strange matter - sure you might be able to produce gravitational waves to protect against them, but then you might have to deal with strange matter AND incoming gravitational waves. With effectively infinite resources and arbitrarily advanced technology, offence always wins in the end.
I'm particularly unsure about whether I'm right about this statement: Even if a superintelligence exists in your solar system, the creation of a superintelligence optimised for total destruction somewhere else in the galaxy will inevitably lead to your destruction.
How bad this would be for the long term potential of sentience is very dependent on the aliens that destroy us. It’s possible that the aliens might be intelligent but not sentient, so if they replaced us that would be a disaster. The aliens could also be tyrannous, or have little respect for other sentient beings, and thus their conquering of our galaxy represents an s-risk. Much ambiguity.
Yeah this took me more than just one afternoon. I really missed the mark on this "exhaustive list of cosmic threats" in 2023, which lists a pathetic 19 threats.
FYI there are like a million caveats and conditions with each thing on the list, so if its confusing I'm very communicative over email or on the forum and I'm happy to answer them if you're confused or interested.
There is so much uncertainty here that these can only really be guesses. I'm happy to update them if other points or new research comes to light. Maybe I'll re-write the post in 2040.
Without a strong presence in the real world, this would leave the civilisation vulnerable to the emergence of alien life elsewhere in the universe though, which is probably an inevitability given enough time.
Once we have this governance system, we probably need to spread throughout the affectable universe as fast as possible to prevent independent alien civilisations from emerging, or gain the ability to collaborate with them on preventing galactic x-risks.
The seminal work on the subject is this paper: Armstrong, Stuart, and Anders Sandberg. "Eternity in six hours: Intergalactic spreading of intelligent life and sharpening the Fermi paradox." Acta Astronautica 89 (2013): 1-13.
Wikipedia: "Mars and Earth during the Cryogenian period may have experienced the opposite of a runaway greenhouse effect: a runaway refrigerator effect. Through this effect, a runaway feedback process may have removed much carbon dioxide and water vapor from the atmosphere and cooled the planet. Water condenses on the surface, leading to carbon dioxide dissolving and chemically binding to minerals. This reduced the greenhouse effect, lowering the temperature and causing more water to condense. The result was lower temperatures, with water being frozen as subsurface permafrost, leaving only a thin atmosphere.[38][39] In addition, ice and snow are far more reflective than open water, with an albedo of 50-70% and 85% respectively. This means that as a planet's temperature decreases and more of its water freezes, its ability to absorb light is reduced, which in turn makes it even colder, creating a positive feedback loop.[40] This effect, combined with the decrease in heat-retaining clouds and vapor, becomes runaway once snow and ice coverage reach a certain threshold (within 30 degrees of the equator), plunging the planet into a stable snowball state.[41][42]"
2
1
