AI Built for the Future Exploration of Space

By Jackson Noel Davies | 14 August 2017
Medium

One of the major problems with space travel is the fact that being able to travel faster than the speed of light (and more importantly surviving it) are going to be nigh on impossible for Humans, based on our understanding of the science so far.

Methods of Human Transgalactic Space Travel

FTL, Hyperspace, Wormholes, Warp Speed, Folding Space, Blink Drives and Particle Translocation are all ideas only theorised in deepest science fiction. Sadly we haven’t found Arrakis yet, or the Spice, let alone traveled there to go and mine it.

As yet, we’ve only been to the moon, and when I say we, I mean the few NASA astronauts that did it on our behalf. We’ve sent out a dozen probes to far flung reaches of the solar system but haven’t produced a vessel that has journeyed beyond our own solar system to another.

Alpha Centauri is our neighbour, 4.24 light years away as the crow flies, or 1.3 parsecs if you prefer parsecs.* Proxima Centauri is the closest star in this system, a bit bigger than Angelina Jolie, or George Clooney. Alpha Centauri, funnily enough, is a trinary system. It has a binary pair of stars (Alpha Centauri A & B) and a solitary red dwarf (Proxima Centauri).

while interstellar travel conjures up all kinds of visions of Faster-Than-Light (FTL) travel, ranging from warp speed and wormholes to jump drives, such theories are either highly speculative (such as the Alcubierre Drive) or entirely the province of science fiction. In all likelihood, any deep space mission will likely take generations to get there, rather than a few days or in an instantaneous flash.

www.universetoday.com — 26th January 2016

The article from universe today looks at our fastest sensible method of travel to Alpha Centauri and by the estimation it would take 87,000 years to travel there at ‘max chat’ or ‘full bore’. That is likened to 2,700 Human generations. An impractical length of time.

Even the most theoretical of theoretical would take under 4 years in the shape of the Alcubierre drive system, a warp drive-esque idea where space is manipulated to allow the vessel to travel quicker by means of a wave form that disrupts the fabric of space without breaking Einstein’s laws. It’s a big stretch in terms of technical capability and last on our list.

Needless to say that any options we have to make such massive crossings are not cost effective right now but it doesn’t mean we can’t start on our way to exploring.

*The closest object in the Alpha Centauri system is the Red Dwarf Proxima. It is difficult to discern any objects closer in that system with our current level of technology. Also of note is that some areas list Alpha Centauri as being 4.37 light years away but that measure is the centre of the system, the binary stars Alpha Centauri A & B.

***

Using AIs to Explore Space For Us

Being weak and fragile, and requiring to procreate and keep generations moving forwards on insanely long space journeys that could all be for naught, it stands to reason that sending AIs to do the looking for us saves both time and blind hope.

In SimCity terms Humanity’s vessels will be flying at ‘turtle speed’ to their destinations, forget ‘Cheetah’, there isn’t even a button for that.

Space travel is going to be a deep cruise, a deep deep cruise, a deep deep deep cruise. You get the idea.

In order to send a new colony ship somewhere, we are going to expend insane amounts of resource in order to maintain a colony that not only sustains but grows with the generational population as it streaks into the stars. There are a whole series of challenges that run hand in hand with maintaining a population off world, and as a Human race, we’d have to figure out a lot of those challenges in order to get started. It isn’t worth it overall. Not at the slow pace we’d be travelling.

For an AI such ideas are different.

***

Benefits of Just Sending an AI to Map a Nearby Star System

So we agree that it is going to take a long time to send an AI. If we sent one tomorrow with our current stage of technology, it would take 87,000 years based on the best we have right now, and that’s just the nearest star system.

I’ve considered 8 benefits but there are bound to be more.

  • Benefit one: Smarter Upgrading

If we were to allow the vessel to be modified by the AI as it learns new ways to maximise the potential of propulsion, it would eventually get faster. If we were able to communicate new designs over the vast distance of space, and provide additional building blocks on the way, or even build said building blocks into the initial design, it is quite possible that we’d be able to speed up the process considerably.

  • Benefit two: Smaller, More Flexible Capacity

We could save masses of space by just providing a simple core for the AI. If we programmed the AI to create machines when it needed them, through a process of modular combinations that remain in storage until needed (a bit like a lego kit), we could allow the AI to perform some complex tasks without needing to worry. It could change the mass and configuration it uses up at will.

  • Benefit three: Simultaneous Deployment & Shared Learning

We could send AIs to multiple star systems simultaneously. Through doing so, we could chain link each AI to learn from the other AI so that we cut the learning curve of all of the AI in one hit. Hive data sharing would be beneficial as those AI that reach destinations quicker than others, start whittling down science, and other AI systems can be used to bolster their knowledge by group think and shared consumption of data (a conference of sorts where consensus is determined).

  • Benefit four: Null State

The AI could simply remain in hibernation until they reach a set point. This would mean that all expended energy would be directed purely at the transportation, none wasted on operations beyond navigation, collision avoidance and propulsion. Even if a null state isn’t preferable, a form of periodical hibernation can be instigated.

  • Benefit five: Endless Number Crunching

Number crunching is something that an AI would relish. As it slowly approaches its destination it can redefine its interpretation of what it sees and how it looks at this data (as its target moves closer into view). It can start to define tactics to effectively determine points of interest to view first and work on methods of opening a line of communication between Sol and the target system. It can devise methods of science that are appropriate to what it can see, rather than what Human scientists speculate to be there.

  • Benefit Six: Neutral First Contact

First contact is de-risked to some degree by the virtue that machines are more logical than organic mis-communications. An AI is the perfect third party and can adapt to all the required handshakes of its new companions without fear that it will cast Humanity in a poor light. As AIs don’t have the same resource requirement as organic life forms, it is likely that little friction will be generated between sentient species and itself. This would be especially useful if the AI is programmed to consider harmony and compromise, over domination and posturing that might come from a more traditional organic interactions. An AI has far less reason to be Combative when all it needs to do is observe from afar, and generally be nosy without interfering.

  • Benefit Seven: Unlimited Exploration

Allowing AI to map our nearby systems saves us the time when we finally get there ourselves. They can log and look for all of the interesting phenomenon that we specify, or that they consider extraordinary by their own evolving definitions. Being non organic, particular radiation zones and more dangerous phenomena can be charted without endangering organic life. AIs and anything AIs can create, will definitely be able to travel further around new systems than we can, and will be able to adapt to worse extremes more rapidly than we can. We would be able to quickly eliminate systems that aren’t good cradles for life, and instead allow the harvest of such systems (if that was deemed ethical).

  • Benefit Eight: Self Proliferation

Self proliferation could be brought into function. When an AI arrives safely at a destination and determines that resources are available, it can begin generating a new vessel based on not only it’s own designs but those from the consensus. It then sends out clones of itself with the mission of establishing research in other nearby systems that have not yet been explored whilst performing its primary mission of exploration. This would significantly cut down time wasted in exploration. It is easier to send a new exploration vessel from a closer point (because you have to factor in less fuel if say we are using Hydrogen based fuel), at the very least we would expend less resources on Earth if we don’t have so many vessels to build in the first place.

***

Other Benefits of Letting AIs Do the Heavy Lifting

Space is a lonely place.

AI are far less likely to get bored and kill themselves than a Human. We can provide an AI with lots of math problems to solve whilst it is on its very slow cruise to the next star system and it will happily calculate this. If we don’t have the energy in reserve in the design of vessel we are sending, the AI will sleep and only wake when it is close enough to the system to start its work.

If we are concerned that a single AI will have problems being alone (a very human concept), we could always provide a conference of AI. 3 or more, like a person with multiple personalities. We could provide each AI with different functions so that they have a unique value to each other, and must agree in majority to perform an action. This may then help preserve existing eco-systems and unusual phenomena which might otherwise be labelled as re-use for further exploration purposes.

The pure benefit of AI is that they are entirely consistent and persistent. Think if Einstein was allowed to live forever. That is your AI. Getting smarter every second, without the human fragility.

***

There Are Drawbacks….

….But I’m going to cover those in a follow up article.

For now, rejoice in the thought that whilst we might not travel to the system first, we’ll know practically everything about it by the time we finally get there.

Let’s put robots in space: Philip T. Metzger at TEDxOrlando

The Future of Asteroid Mining

How humans could evolve to survive in space | Lisa Nip

Be sure to ‘like’ us on Facebook

LEAVE A REPLY

Please enter your comment!
Please enter your name here