Project Hyperion: Will a “Seed Ship” Give Birth to The “Star Child”?

posted by Andreas Hein on June 1, 2012

written by Christian Bühler


Artist rendition of an artificial womb


A new civilization is born

Imagine a planet in an extrasolar system far away. It looks a little bit like our earth. You see blue oceans and huge continents. In the next moment a star ship enters the silence sending out a fleet of robot ships. You watch them beginning to build bases on the surface and in the orbit of the planet. As the constructions grow bigger and bigger you decide to find out more about the ship. You fly slowly around the ship watching the two powerful engines and the shimmering hull, still not knowing why it was sent. You reach the front end wondering about the huge antennas receiving signals from somewhere far away. Suddenly you hear some noise from inside the ship. After listening for a while you slowly recognize it must be a baby which has began to cry. What happened? The star ship you have discovered is a seed ship bringing a new civilization to the world beneath you.

You do not know what a seed ship is. Okay, let me explain. A seed ship is an interstellar spacecraft which is able to transport embryos or gametes over long distances. In order to keep them save and healthy it acts most of its flight like a huge fridge, because the unborn humans can only survive such long travel times in a frozen state. Close to the target system this “fridge” is transformed into some kind of brood chamber. Next the chamber begins to grow up the gametes or embryos to babies. After a few years when they have become older, the settlement can start. The two main questions that appear in this scenario are about transportation and how to parent and educate the children. The answers especially to the second question are quite difficult and controversial as you will see below.


A closer look at the transportation options

Right now, you might be able to imagine how the whole scenario of a seed ship settlement might look like. What I want to do here is to go further into detail. So let´s start at the beginning. What are the different ways to transport unborn humans to the stars?

–    The first option is to send frozen embryos to the target system

–    The second option goes one step back and you send donated egg and sperm cells

–    The last option is to send only genetic information and produce the required egg and sperm cells at the planet

      you want to settle.

The last option might be not feasible at the moment, because there is no system such as a biosequencer that can create the sperm or egg cells. So let’s have a closer look at the two remaining possibilities.

Both options for frozen embryos and frozen sperm and egg cells may be no problem due to the action of freezing. This is already state of the art. But an interstellar trip might last over more than a hundred years and we do not know if the cells or the embryos might take serious damage over such a long period of time. Today we only know that children were raised of embryos or gametes which are not stored longer than 20 years. We need a few more long term studies to confirm that freezing is no problem. These studies will cause ethical difficulties because you would need human embryos as test objects.

Another requirement for the freezing options is that there must be some kind of artificial uterus to grow the embryos or gametes. An artificial uterus is a machine which makes it possible to grow embryos or gametes outside of a woman’s body. These machines might become necessary because there are no adult humans on a seed ship. If we go back to the brood chamber example it makes no sense to carry eggs to a planet if there is no possibility to hatch them. An artificial uterus does not yet exist for humans, but Nick Otway invented one for sharks. Due to this fact, it can be assumed that during the next decades this technology might also become feasible for humans.

Raising the babies

Now let´s go one step further. The “new settlers” are transported safely to the target system and are now grown up in the artificial wombs. But this leads us to the next problem. It is necessary to parent those children and provide proper education in every field that is essential for the establishment of a new settlement.

One option might be to provide the required education by robots. But there are several disadvantages to “robot nannies”. One problem is the need of an advanced artificial intelligence for robots in order to parent and educate children. That might not be such a big issue because before we are able to send an interstellar spacecraft to another star, this technology might already be invented, as the spacecraft will require advanced artificial intelligence. Whether or not this type of artificial intelligence can be modified to satisfy the requirements of a “robot nanny” is an open question. The greater trouble is to test if it is really possible to use “robot nannies”. We cannot grow up thousands of babies at the target planet and hope everything will work out, without having done several tests beforehand. This causes a big ethical problem. We cannot do these tests on animals like we do for medicine. We must perform the tests on human children even if we do not know the results. Our society and I think almost everyone else does not agree with using children as laboratory rats.

Solving the “robot nanny” problem

So we still have the problem how to parent and educate children at their new home. One possibility is to form “teacher ships”. These ships are a combination of seed ships with sleeper ships. These are interstellar spacecraft which transport hibernating people to the target planet. If you want to find out more about sleeper ships and hibernation please feel free to read Daniel’s blog article. For example, let’s assume that the sleeper ship part of the teacher ship is a habitat for about 25.000 people with a hibernation chamber in which at the beginning only 1.000 adult crew members hibernate. After 25 years of journey these adults are woken up. At the same time 4.000 babies are grown up in the artificial wombs. Now the adults can parent and educate them. This education period will last about another 20-25 years. At the end of this wake phase there are over 5.000 well educated adults. Everyone is send back to hibernation again for 25 years. In the next step 20.000 babies are grown up. At the same time the adults are woken up and able to parent and educate the newborns. After the second education phase, which will also last for 20-25 years, about 25.000 educated adults live in the spacecraft´s habitat. At this point of the journey the teacher ship enters the target system. After reaching the target planet and the building of orbital bases or bases on the surface, more and more embryos or gametes can be grown up until the whole civilization is established. Sequencing the hibernation and education phases in a manner that only let’s a small part of the population awake might allows to drastically reduce the habitat size and life-support system requirements.

A second option might be to combine a world ship with a seed ship. In this case the seed ship part of the world ship is only an additional freezing chamber which is used to transport the unborn babies. The world ship brings e.g. 25.000 people to the target planet and after building orbital bases or bases on the surface the embryos or gametes can be grown up. The 25.000 adults which live on the world ship than can parent and educate the newborns. That means 25 years later we have a population of about 125.000 well educated people living at the target planet if we use the same number of babies per adult as it was used for the teacher ship.

These options require much more mass than a pure seed ship, but if we compare it to other possibilities like sleeper ships or even world ships without gametes or embryos with the same mass we can send much more people to another planet, as we carry the rest of the whole civilization as gametes or embryos in the “backpack”. This ensures a higher genetic variability over longer periods of time after colonization. We come to the conclusion that seed ships or modified seed ships seem to be a good option for manned interstellar spaceflight, if associated technical problems of cryogenics, artificial wombs and education are solved. Nevertheless the associated ethical problems are obvious and further research has to be done in every subarea before a technically and ethically viable solution can be proposed.


Be Sociable, Share!

17 Responses to Project Hyperion: Will a “Seed Ship” Give Birth to The “Star Child”?

  1. JohnHunt says:


    A couple of years ago, I wrote up a fairly comprehensive but concise description for such a seed ship mission. In case you haven’t read it, it is at:

    In it, I reference work done by Hung-Ching Liu at Cornell University in which she was able to raise mouse pups to the equivalent of our 31 weeks gestation using an artificial uterus. I believe that ectogenesis of humans is well within ethical reach by mid-century.

    The obvious thing which jumped out to me in this article is this question, “If we can send a teacher ship, why do we need a seed ship? Couldn’t those teachers reproduce naturally or by in vitro fertilization”?

    The only reason to even consider a teacher ship is if a robonanny is not possible by the time of launch.

    The common (mistaken IMO) assumption is that you would need true artificial intelligence in order to “teach”, or more importantly, raise, children. I can understand why people would assume this. Parenting is a complex thing, no doubt. But so is verbally communicating with people. And yet Siri illustrates strides along those lines. Imagine what Siri will be (by way of course) by mid-century. You also may not be aware of the leading contender for the Turing Test. You might want to Google, “cleverbot” to see the state of the art. My main point here is that there is a clear path to passing the Turing Test without having to develop true AI. We can get there with an expert system alone. As for the physical behavior of the android parents, realistic looking and moving androids already exist. Some of the behavior can simply be recorded for replay just like much of the conversation. Again, please read my paper. I have already addressed the feasibility of these things.

    From my perspective, the main advantage of a seed ship over any other manned variant is that it would be lower mass than the other options. If launched with beamed propulsion, and especially if we were to allow a long travel time, the power and infrastructure needed to launch would be much less than any other interstellar mission. An earlier “manned” launch can buy a life insurance policy for the human species in a way that no other interstellar mission can.

    Regarding the ethical considerations, I have written a supplementary document addressing the various ethical issues. That is at:

    In particular, please look at my comments about how android parents could be ethically developed. Every day children are used as research subjects and I suggest a cautious way in which the android parents could be developed without harming the child subjects.

    • Christian Bühler says:

      Thank you for your reply John.
      I read your articles with a great deal of interest. I understand your argumentation that true artificial intelligence is not necessary to educate and parent children. In my opinion the problem on this point is that you imply that the children are able to speak to communicate with the androids. If this is possible I agree with you that the hole thing can be done without a true AI, by using something like the “cleverbot”. But babies are not able to speak. Most of their “communication” is done by crying. The clear understanding what they want to express is sometimes even hard for human parents. So I do not think the parenting for babies can be done without a true AI or human parents.
      The second thing you mentioned was that there is no problem to use androids as parents. I do not think they can really replace human parents. In my opinion the use of androids would cause problems like hospitalism. This disease occurs if children especially babies suffer from a lack of “nest warmth”, which I think can only be given by human parents. For that reason I hold the view that androids can only be used to support the parenting.

      • JohnHunt says:

        Very interesting. No one has previously brought up the issue of the parenting of infants.

        The most common cries of an infant are for a limited number of reasons: hunger, tired, wet, discomfort, pain. Some of these cries have unique characteristics (e.g. intensity, pitch, and duration) which a computer could identify. As with a new parent, the robot could attempt to alleviate the cause. If one does not work then other approaches can be tried. Please bear in mind, automated parenting would only be initiated if humanity had gone extinct. Although we would attempt to provide the children with the best upbringing possible, survival, not perfection is the goal.

        As for hospitalism. This most often occurs when infants are neglected in incubators. Androids can be programmed to provide as much tactile and auditory stimulation as we care to give. For example, people could be recorded playing “peek-a-boo” which could then be replayed on an android. Or, the android could be programmed to play, “Here comes the choo-choo train”:

        Indeed, a whole repertoire of actions and speech could be recorded and replayed according to rules.

        Just look at this Disney animatronic video. Although other androids have realistic skin and facial gestures, this one has a recorded personality which is indistinguishable from a real person:

        Let’s remember that there is anywhere from 50-88 years of technologic development before launch. Much of the android technology is at relatively high TRL levels as evidenced by the videos. I honestly doubt that android parenting will be the hardest part of the mission.

    • Alex says:

      Hey Everybody, just wanted to let you know that this article inspired me for my the last Episode in my Sci fi series :
      Many thanks

  2. My difficulty with these concepts is that they require civilisation to be compressed down to a very small, non-functional payload, which then has to be reconstructed automatically on arrival. Unless a very closely Earth-like exoplanet is found, this would first have to be demonstrated within the Solar System, together with demonstrations of raising of healthy children after storage as embryos over century-long or millennial timescales. Closely Earth-like exoplanets will only exist at a tiny minority of destination stars, and despite “Avatar”, I think we should plan on the assumption that the first several interstellar flights will not have such a planet at their destination.

    The space colony mode of life which will be demanded at most colonisable stars requires prior large-scale space colonisation within the Solar System, which provides both the experience base and the size of economy needed for worldships occupied by people living normal lives. It is interesting to contemplate alternatives, but to my mind sleeper ships and embryo ships add extra technical problems for little gain.

    • JohnHunt says:

      Hi Stephen.  Thank you for your thoughtful reply.

      The potentially dangerous assumption is that humanity faces no existential threats prior to the normal colonization of the solar system, the development of He-3 propulsion, world ships, etc.  But reality is not obligated to conform to our hopes for the future.  In fact, I find it apparent that technology posing an existential threat is steadily advancing along at least four fronts with their maturation likely to occur before the end of this century.  Please consider what will be possible after the next 88 years of development of self-replicating chemicals, biotech, seed AI, and nanotech.  Again, will you please spend some time thinking about this and it’s implications for the first interstellar mission?  I don’t advocate embryo space colonization because it is easy but because it may be necessary.

      > require civilisation to be compressed down

      Civilization results from initial colonists, sufficient resources, and information to eventually rebuild civilization.  I believe those can be compressed to small size.  True, reconstituting them is the difficult part.

      > to a very small, non-functional payload

      Yes, in order to minimize mass hence beamed energy hence space infrastructure hence time to launch.

      > a very closely Earth-like exoplanet

      No.  I agree that a “closely Earth-like exoplanet” is probably not near.  I don’t presume 1 gee, oxygen atmosphere, warm oceans, etc.  But I do presume something between a Mars and 1.5 x Earth not necessarily within the habitable zone.  Our exocivilization will be more like a Mars colony than Earth-like.  We will need to create our own habitats.

      > demonstrated within the Solar System…raising of healthy children

      Why not save money and demonstrate it on the Earth such as at some arctic station?

      Please read my ethics article.  There is a way of doing R&D with automated childrearing that is ethical and can be done on Earth.

      > after storage as embryos over century-long or millennial timescales. 

      We can expose frozen embryos to the same thermal and radiation stress of the full mission and then check their viability.  Starting with 1970’s frozen embryos, we can see if there is a decreasing viability trend.  I believe that current research indicates there is not.

      > but to my mind sleeper ships and embryo ships add extra technical problems for little gain.

      Unless the gain is the survival of human civilization.  In which case, waiting for alternatives is to risk losing absolutely everything, right?

      The Embryo Space Colonization to Avoid Possible Extinction (ESCAPE) Mission is a unique interstellar mission which fills a need (avoiding near-term extinction) in a way that no other interstellar missions can.  I believe that it should be one of the highlighted interstellar missions to receive design attention.

      • My suggestion would be to refrain from speculating about existential threats to civilisation, and focus on developing a better future for all, whether on Earth, elsewhere in the Solar System, or diversified on an interstellar scale. We can get to the stars without needing to rely on doomsday arguments. And if we expect public support for our plans, we must show the public how they benefit too!

        • JohnHunt says:

          Self-replicating technologies will likely mature roughly in the latter half of this century. I fear that a positive but indirect approach on our part will only delay the design and development of the only true interstellar mission able to purchase this life insurance for humanity within this century.

  3. Adam Crowl says:

    Of course there’s always the “reconstituted astronaut” option – transport the cells of the crew in a form we know we can freeze – as cells. Then reconstitute/rebuild as desired at the destination via organ-printers or whatever the equivalent is. Of course I’m not saying we dissolve a living person. Instead we make a body mass of their stem-cells, scan them in sufficient detail, then duplicate them at the destination – producing a “clone” but skipping the slow process of normal development. Of course it might be psycho-biological fantasy, but what if it’s not?

  4. I guess that essentially sending canned humans into interstellar space with no prior consent on their part for the mission would be beyond acceptable ethical limits. The public will hardly support such a thing.

    • JohnHunt says:

      In vitro fertilization from frozen embryos occurs all the time without much ethical concerns and without anyone objecting to the frozen embryos as being “canned humans”.

      The issue isn’t so much that the humans are stored in the form of frozen embryos. Rather the issue which you are probably getting at us that the children from those embryos would necessarily be raised by androids, in a setting of risk, and reduced social opportunity — all of these circumstances without, by necessity, the consent of those children.

      Several things. First, by necessity, parents make all sorts of decisions which affects their children without their consent. Second, the raising by androids would not proceed unless humanity had apparently gone extinct. This would create a strong ethical rational for why the process must proceed. Third, as robotics and expert systems advance, we can steadily improve the degree to which androids accurately mimic real parenting. By the time of launch, society would know for a fact the typical outcomes of android parenting. Their views would be based upon the results of the android parenting R&D program rather than just a theoretical gut feel about what might happen.

      If all of us truly believed that humanity within the solar system was going to go extinct by a certain date (let’s say 2067), I think that our discussion would be very different. Ethical concerns would largely be placed aside as the necessity of a “manned” interstellar mission would be apparent. Thoughts of a science probe within the lifetimes of participating scientists would be placed aside as irrelevant. “Ad astra increments” would also be placed aside and replaced by a rush to achieve a specific, practical objective.

      The fact is that none of us knows whether or not humanity faces an existential risk within this century. Our accelerating technology and Ferni’s Paradox suggests that we might want to take this possibility seriously. At a minimum, the ESCAPE Mission should receive similar technical investigation that other interstellar concepts are receiving just in case the risks are real.

  5. tristanvdb says:

    Only one thing to say: Songs of Distant Earth by Arthur C. Clarke.

    In fact, I will add that Clarke in this novel address the only kind of doomsday that will require evasion from the solar system: its destruction…

    For the issue of raising children using android, he have a simple solution: the first generation will be traumatized (I don’t remember how he exactly said it, but it was good).
    To develop on the first generation idea, one should start with a small number of human raise not to become functional adults (by this I mean able to leave in our society) but to be correct parents (possibly some genetic tweak: low IQ, high attachment, …) and produce more functionnal generation after…


  6. Zanstel says:

    I prefer a two stages journey. One, using a big world ship completely “frozen” even no gravity or atmosphere, to allow a simple journey with “slow” speed, perhaps with a travel time of centuries.
    Later, a very fast ship, that “burn itself” (like a multistage “rocket”, where all matter, even the walls, are in some point converted into plasma propellent, so the ship shrink during the travel), to reach speeds near speed of light with great acceleration (1g?= no artificial gravity), with little humans (a hundred perhaps), hybernated or living in a small space) to reach the target in little years.
    With fusion, the “little” ISP needs a huge ratio of mass on origin vs destination, but the small crew allow it.
    So in destination, they arrives to the world ship, put in into work, rotation, “atmosphere”… and uses artificial wombs to grow childs with big ratios ( 1:10 per crew). In three generations, the world ship could be fill, study the target deeply and begin to work to replicate the world ships with native matter (mainly using asteroids and comets).

  7. Pingback: Interstellar Travel Here and Now @ Helian Unbound

  8. Robert Lucas says:

    You’d do better with suspended animation (is that right what I’ve just written?) People want to wake up there don’t they. A ship of people having babies would be better probably.

  9. Autumn Kalquist says:

    But *why* would we do this?

    The only way I can see us building ships like these is if living humans wanted a chance to explore, or if the Earth became uninhabitable. In that case, you’d always have adult humans aboard.

  10. Stefan says:

    It occurred to me just yesterday that space travel is the best use scenario for an AI that would eventually surpass our level of intelligence. Such an AI can be EDUCATED to seed life throughout the Universe, either by spreading OUR biome and genome or, in case it finds a planet that has the right temperature and atmosphere, but not the right elements, engineer molecules and eventually living organisms based on whatever it finds on the spot. Imagine such thing being done on Titan, which has plenty fluid material. If it seems too sci-fi, well, it is. But so were many things 100 years ago. An AI that has an extensive database of all human knowledge at it’s disposal and it’s capable of learning from it AND furthering research and knowledge, should be capable of coming up with “functional” molecules based on a chemistry that’s alien to life as we know it, provided it has the right tools to work with.
    Ultimately this would be the absolute selfless thing that humans can do with their knowledge: start up life, regardless of being us, or like us, or completely alien. Our genes command us to push forward, to strive and survive, but if we ever face an ultimate demise, incapable of saving ourselves, at least we can preserve life by attempting to create it somewhere else, remotely, in another time, without seeking benefits or rewards.

Leave a Reply

Your email address will not be published. Required fields are marked *