Exactly How A.I. Could Be the First to Find Extraterrestrial Life

A critical aspect of NASA’s mission to explore our solar system and beyond is artificial intelligence, especially in uncrewed probes that venture into deep space beyond our solar system. So it stands to reason that if there is extraterrestrial life sitting out there in the solar system, an artificially intelligent system will find aliens before humans do.

NASA expects this too outcome, of course. It’s why it has an entire division dedicated to developing and advancing A.I. systems relevant to space exploration. Steve Chien is the technical group supervisor for the Artificial Intelligence Group at NASA’s Jet Propulsion Laboratory in Pasadena, California and is at the forefront of this project. Chien’s group is tasked with making sure whatever probe that ventures into regions around Jupiter, Saturn or beyond has the ability to assess its environment, its changing circumstances, and adapt accordingly.

“It’s just like if we’re out running errands, we’re dynamically calculating things, estimating things all the time, and sometimes things take longer, sometimes things take shorter,” Chien tells Inverse. He cites work on the Mars rover programs — the next iteration which will be the Mars 2020 rover — as an example of how NASA is trying to use A.I. to give its technologies more agency to figure out on their own which things are more important to investigate further using more time and resources, versus other features. “And so we’re trying to give the rover some of the ability to do that.”

The Mars 2020 rover, which will be explicitly tasked with conducting astrobiological research on the red planet, could very well be the first robot to find signs of alien life. But the Mars 2020 rover is just assisted by A.I., and wouldn’t operate strictly as an autonomous probe with its own independence.

The incentive for getting a probe to operate fully autonomously is because the farther out the spacecraft gets, the larger the lag time in communications. A whole lot could happen in the span of those couple hours, and ground doesn’t have the luxury of sitting by at every given moment to provide instructions. A probe will have to think for itself.

NASA will get closer to full A.I. with something like the Europa Clipper flyby mission. The main challenge for looking for life on Europa is actually getting to Europa safely. The Jovian environment is a radiation-filled hellhole. Chien explains that the radiation forces any spacecraft to reset its processors, about five times per flyby.

The mission is looking to conduct something like 40 to 50 flybys.

“We want this autonomous system to be smart enough that when you have this reset,” says Chien, “it recovers the system and goes back to the signs that was originally planned as quickly as possible. That’s not easy, because the plan to observe things was very carefully planned out by people on the ground, a lot of scientists negotiated, and it’s fairly complicated. So getting the system to reset and do it safely and appropriately in all those different cases, regardless of when the upset happens, is very challenging for an A.I. system.”

Chien explains that NASA scientists haven’t found a solution yet, but its these kind of technical problems that an autonomous spacecraft needs to be able to troubleshoot on its own.

And if we decide to land something on Europa or another potentially habitable ocean moon and dig through the ice to search for life? Well, we’ll need a robot that can autonomously navigate through an environment which we’ve never seen before.

Fully-autonomous systems are the goal for NASA. Chien explains that while NASA has to start working on something like that, there are plenty of visionary ideas being batted around. One is to develop a a type of technology that could turn near Earth objects (like asteroids) into autonomous spacecraft that can be sent off into deeper reaches of space to conduct observations (sort of an inverse concept of Made in Space’s vision to make asteroid automata that will deliver resources to Earth). Those craft could very well stumble on extraterrestrial life in the far reaches of space as well.

But the biggest promise for an A.I. spacecraft to find aliens is in something more interstellar. In fact, if there’s life in other star systems, an autonomous spacecraft would be basically necessary to find it.

“It’s just very hard to get to another star, let’s say, four-and-a-half, five light years away,” says Chien, referencing nearby star Proxima Centauri and the Alpha Centauri system — both of which are thought to potentially possess habitable worlds.

There are ideas for how to study those systems relatively soon, but they neglect to solve the braking issue, which makes it impossible to do robust scientific study. To solve the braking system, however, you need A.I.

“[The spacecraft] needs to be self-healing, self-aware, able to manage all these things, manage problems that come up,” says Chien. Once it’s beyond a certain point, “it’s pretty much on its own. It has to say, ‘Where are the planets?’ Does it get into orbit around one of them? Which one? What measurements does it take? What data does it send back? Keep in mind, it’s not so easy to send back data from that distance.

“And it’s just gonna send the data, let’s suppose it lasts for 10 years there,” says Chien. “That means that we’ll be able to communicate with it and we’ll send back its first data, and when we reply with a message, it’ll be getting there around the end of the mission’s lifetime.”

That means a spacecraft strewing around Alpha Centauri will basically be on its own for the entire mission. It will need to decide on its own precisely what’s worth investigating and what isn’t. And it will need to decide how to act accordingly if it discovers signs of alien life on a distant planet.

That’s perhaps one of the toughest things for NASA and other space researchers to figure out — not just in terms of technical obstacles, but also just in terms of the formulating the best scientific and ethical approach. Should a spacecraft attempt to make direct contact? Should it make its presence known? Should it try to obtain a sample? How can it limit any negative impact it may have on the environment of the star system?

Those are all insanely large questions to answer, but the one thing that’s clear is that is going to take the most cautious approach possible, especially for A.I.-mediated systems. “[NASA is] very risk-averse, and it’s perfectly logical,” says Chien. “In fact, I support it. There are so many challenges about flying a successful space mission. You have to launch, you have to land, you have to drive around. There’s a thousand components, or a million components, and if a few of them go wrong, you lose your mission.”

So while Hollywood loves to portray robots as crazed overlords, chances are the ones that finally discover whether or not we are alone in the universe will be some of the safest systems ever made.