Alexa in space? Why free-floating AI friends are in our interplanetary future

This post is part of Science of Sci-Fi, Mashable's ongoing series dissecting the science (or lack of science) in our favorite sci-fi movies, TV shows, and books.


Not many American nerds these days know about a golden age sci-fi writer called Edmond Hamilton. If they do, it's because of his Star Wars connections: Hamilton was the husband of Leigh Brackett, space opera queen and author of the first draft of The Empire Strikes Back; he also happened to be the first guy to use a "laser sword" anywhere in fiction. 

But German nerds tend to remember Hamilton for something completely different, and the future may do, too. Because he didn't just invent lightsabers. He also invented free-floating space brains. And in so doing, inspired a futuristic kind of Alexa built along those lines — an intelligent assistant that recently had its first orbital test run, and now seems poised to join our nonfiction interplanetary travels. 

Here's how that happened. In the 1950s, Hamilton wrote a series of pulp stories about a character called Captain Future. The titular, predictably square-jawed hero was rather less interesting than his mentor, the encased floating brain of an elder scientist named Professor Simon Wright. (Think Obi-Wan Kenobi, if he were inside that hovering training remote Luke Skywalker had to fight.) 

In the 1970s and 80s, a modernized anime version of Captain Future was a hit in many countries around the world, Germany most of all. German kids thrilled to the adventures of the Captain and Simon on screen, then turned to the Edmond Hamilton books the way we burrow into Harry Potter. 

Then the Captain Future nerds grew up, and some of them went to work for Airbus and IBM and DLR (the German NASA). There, these fans built an actual talking, emotion-sensing, drone-like floating brain, and blasted it into orbit with real-life Captain Futures — astronauts aboard the International Space Station. 

Its name? CIMON (soft "c", please) which in theory stands for Crew Interactive Mobile Companion. But German space nerds immediately got the reference. 

"My first impression was, 'wow, that is really a Captain Future scenario,'" says Volker Schmid, a mission manager at DLR. He's talking about the moment in 2015, at a conference called Humans in Space, when he came across a team from Airbus presenting a paper on how they could build the first artificial intelligence for space missions. In zero gravity, it would be easy to give this "brain" autonomy to move around, using built-in air vents and fans to propel itself, rather like a drone.

SEE ALSO: The 10 alien species we'd most like to invade Earth right now

Schmid was a Captain Future fan who happened to be trawling for experimental projects he could send to the International Space Station alongside German astronaut Alexander Gerst. He commissioned a team of 50 people at Airbus and elsewhere to build the brain, fast-tracked it so it would be ready for launch in just two years, in half the time such a project might normally take. 

He also made sure the name was in "homage" to his childhood hero. That, he hoped, would fend off the inevitable comparisons to HAL, the malevolent, murderous space AI in 2001: A Space Odyssey.  

There was a serious side to the notion of floating  brains, of course. Future astronauts will need a resource like CIMON the further away from Earth they get. On long trips to Mars and beyond, out of range of meaningful conversation with mission control, the knowledge base of a voice-activated system — medical knowledge, technical knowledge, even ethical knowledge — could be invaluable.   

"You can't have every skill in your crew, it's impossible," Schmid says. "You can't keep 200 experiments in your mind for a Mars mission. We see humans and robots as a team." 

Simon Says

You may have seen Simon — sorry, CIMON — in action already. A video of the robot interacting with Gerst led to some bemused news stories, thanks to  fascinating glitches in the voice assistant software. Given the fast-tracked nature of the project, there wasn't much time to train Gerst in advance, so his time with CIMON was very much the space equivalent of doing it live.

CIMON obediently answered Gerst's requests for "small talk" about space, his command to turn 90 degrees on its own, and provide details on a space station science experiment. But after Gerst asked for his favorite song — which turned out to be "Man-Machine" by Kraftwerk — CIMON simultaneously switched to video-streaming mode and became stuck on being a DJ.

"I like music you can dance to!" enthused the floating brain as Gerst floundered, tried to get CIMON to exit music mode. "Don't you like it here with me?" 

A series of cascading misunderstandings, of the kind familiar to anyone with Alexa or Siri or Google in their home, followed. Gerst forgot that he was supposed to give the AI commands using its name. CIMON started moving through the air of its own accord, told Gerst not to be "mean," then started (perhaps passive-aggressively?) insisting it must be time for the humans' next meal. 

"It's not spooky stuff," insists Matthias Biniok, IBM's lead Watson architect in Germany (currently CIMON uses a satellite internet link to Watson, Big Blue's question-answering business AI system). "CIMON misinterpreted an input that Gerst was giving to the ground station. He heard background noises as well, and he interpreted that as a kind of complaint, and that's why he said 'don't be mean.'" 

Getting hung up on music and meanness is not exactly HAL 9000 levels of willfulness — although it is fitting that IBM, which provided Arthur C. Clarke and Stanley Kubrick with the inspiration for HAL in the first place (move each letter one space forward in the alphabet), is involved. CIMON was much more cooperative in subsequent experiments: It and Gerst worked on a science experiment involving crystals, and tried solving a Rubik's Cube together.

Still, a robot going haywire in space is one of our oldest science fiction tropes — and as much as it cloaks itself in a Captain Future vibe, CIMON can't help but remind us of that. Our biggest barrier to AI in space may be the uncanny valley

For example, that face on the floating ball is designed to comfort the crew. But it also makes CIMON look uncomfortably similar to Gerty, the emoticon-faced robot assistant in the 2009 movie by Duncan Jones, Moon. (Gerty's creepy voice, given the fact that it's Kevin Spacey, now sounds even creepier in hindsight.) 

In theory, assistants like Gerty or CIMON could help astronauts overcome the psychological problems associated with inherently lonely space missions. For now, at least, they're still a little too creeped out by the idea of being diagnosed by a floating brain. 

"We built into CIMON a [voice] tone analyzer, a kind of emotion recognition," Biniok says. "But when we told the astronauts [on the ISS], they said 'well, we really don't want our emotions analyzed.' He laughs. "Maybe with the next [European Space Agency] astronaut, we'll try to activate it." 

Gerst won't get to be that astronaut, since he is heading home for Christmas. The next opportunity to test the floating brain will come in the summer of 2019, when Luca Parmitano becomes the first Italian commander of the ISS. 

But the team behind CIMON is already thinking beyond that. They're thinking about the years and decades ahead, when CIMON has speedier chips and enough on-board terabytes that it doesn't need the internet link any more, because it can store pretty much all the internet inside it. (Okay, maybe not all the cat videos.) 

They're talking about crew members using CIMON or its voice-assistant descendants as a private video diary, or a friend that can open up communication channels with families back on Earth with the same ease that Siri makes calls for us today. 

They're also talking about Earthbound technology that could improve because of the investment in CIMON, such as friendly medically-trained AI for retirement homes.

"There's a lot of work to do to make it more comfortable," says Schmid, the mission manager at DLR. "We need to keep control at every stage, and avoid any developments that will make artificial intelligence [act] against human life." 

That may sound a little famous-last-words-ish, and more sinister than Schmid intended, so let's end on this note. AI assistants on space missions are an inevitable development because we can be much more effective with an external brain, as any smartphone owner knows. And heading to Mars in a tin can is the kind of situation where we need to operate at maximum effectiveness for months at a time. 

But benign AI systems like CIMON are unlikely to have control over the whole ship the way HAL did. The most likely unintentional results of its presence are the same as with Alexa or Siri in your living room: hilarious mishearings and accidental music. 

CIMON can't kill you by locking the pod bay doors. The worst this space ball could do if hacked by some evil entity is to try smacking into you with its flat end, repeatedly. 

And that's where you could use the most essential weapon in humanity's potential arsenal against CIMON. It's a weapon that has been deliberately built in since the very beginning of the project, one that will be essential to every AI descendent going forward, one that even Captain Future didn't have with his Simon. 

It's called the off switch. 

WATCH: Humans have been tossing junk into space for years. This net could help clean it up.

Https%3a%2f%2fblueprint api production.s3.amazonaws.com%2fuploads%2fvideo uploaders%2fdistribution thumb%2fimage%2f86797%2fe4e049ee d059 4bb3 be46 3b10bd9ae88d
Https%3a%2f%2fblueprint api production.s3.amazonaws.com%2fuploads%2fvideo uploaders%2fdistribution thumb%2fimage%2f86797%2fe4e049ee d059 4bb3 be46 3b10bd9ae88d