Inventing AI

Christian Harris
2 min readFeb 17, 2021

Now is an exhilarating time to be a fan of space. Thanks to NASA and its embrace of private aerospace companies, the world is entering into the next big space venture. Companies like SpaceX, Blue Origin, and Virgin Galactic are taking it upon themselves to be innovators in the industry with the hope that commercialization will provide profit. The ability to profit from space travel is something utterly unimaginable for the government-funded programs like NASA and the ESA. However, commercialization of low earth orbit is on the horizon and the commercialization of another planet (Mars) within the next half-century no longer seems impossible.

Flying in space is fundamentally different from flight on earth. Instead of flying in straight lines, space travel is all about orbits. Because of this, there are many possible paths a spacecraft could take while traveling in space. If I were to create a form of artificial intelligence, I’d want to help solve this problem. With a spaceflight trajectory optimizer, any desired flight plan could be optimized for whatever parameter seemed fit. For example, interplanetary space travel takes a lot of time and a lot of fuel ($$$). If sending a lander to Mars, the time it takes to get there is less important because there are no life support systems to account for. Conversely, when sending humans to Mars, the amount of time they are in the spacecraft is an important consideration for any mission designer. A trajectory optimizer would allow for the specific constraints of any mission to be optimized and therefore minimized, allowing for more successful missions. This type of optimizer already exists -developed by NASA- and is known as GMAT, the General Mission Analysis Tool.

There are certainly ethical considerations with such a powerful piece of AI. A spacecraft is nothing but a (huge) rocket and space travel is nothing but very precise control of this rocket. It is important to monitor who has access to such a program because it could find an optimal way of getting a rocket to Mars just as easily as it could find an optimal way to send a rocket from one country to another, with disturbingly precise accuracy.

https://opensource.gsfc.nasa.gov/projects/GMAT/index.php

--

--