Space is the domain of robots. NASA is about to land the semi-autonomous robot Curiosity on Mars within the next few days, where it joins its two less-sophisticated robot brethren, Spirit and Opportunity. There's a good reason why these rovers are the first Earthlings first to set foot — or rather, tire treads — on Mars. Even the simplest robot can survive in space better than a human can.
As we program more and more of our smart machines to explore space, we might discover a lot more than microbial life in the waters of Europa. Instead, says celebrated science historian Richard Rhodes (author of The Making of the Atomic Bomb), we might unwittingly create the first sentient robots.
At SETICon in late June, Rhodes was part of a panel about the 100 Year Starship, a project to design an interstellar ship over the next century. Former astronaut Mae Jemison heads the 100 Year Starship, and explained that part of her plan is to develop important new technologies along the way to the distant goal of a ship that can travel to another star system. Rhodes responded by speculating about what kinds of propulsion technologies you'd need to develop for such a ship, then he commented:
What's interesting are the unintended consequences. I think the intelligence of the robotics is one unintended consequence. There might come a moment when all these intelligent systems that we've been constructing insist on having human rights [and] an equal share in the process. Most likely, the life we'll find in advanced systems will be machines, not something as squishy and difficult as biological organisms.
In this way, Rhodes echoes the prophesy we find in a lot of science fiction — from 2001 to Charles Stross' novel Saturn's Children — where robots achieve sentience in space. This is also a component of the final, heartbreaking scene in Blade Runner, where the escaped replicant Roy recalls his life in space before powering down forever.
But this isn't just a science fictional speculation. Indeed, it's quite possible that our space robots could become the first artificial intelligences. That's because we're programming them to be as autonomous as possible. In space, it's impossible to control a robot remotely all the time. There are time delays, of course, but also line-of-sight problems too. Curiosity may be landing on Mars during a momentary communications blackout, for example, simply because our orbital satellite won't be in the right position to relay a signal back to Earth.
As SETI planetary scientist Nathalie Cabrol put it to me, her job while planning a potential mission to Jupiter's moon Europa is twofold. First, she needs to study Europa-like environments on Earth to get a sense of what to expect; and second, she needs to design robots that can make autonomous decisions about what to study when they drill down into Europa's ice-covered oceans. Essentially, she said, we need robot scientists who can figure out the baseline for what's normal on Europa, and then decide what stands out as worthy of further investigation. Because the time lag between the robots and Earth will be considerable, a lot of the decision-making will be up to the robot.
Ideally, we'd want humans to explore space so that they could make decisions on the fly when they encountered the inevitable surprises. But humans won't be space-ready for a long time. In the meantime, we're designing robots to explore space in our stead. Perhaps, as Rhodes suggests, these replacement humans, these tools, may turn out to be a lot more human than we ever expected.
Photo via NASA/Planetary Robotics Laboratory