If watching sci-fi has taught us anything, it is beware of intelligent robots. From Terminator to Star Wars, when robots become aware humans are in trouble. Elon Musk and a group of conscientious investors are working to make sure that Artificial Intelligence doesn’t reach a dire threat level. Otherwise we may have to also invent a time machine to return to this era and proceed more cautiously.
Although it’s hard to imagine how a robotic arm designed to assemble car parts could become a rampaging menace, it is not hard to imagine Japanese robot bear nurses pulling the plug on a patient due to an overly aggressive computer code. Instead of the irrational fear of Obamacare’s death panels, we could experience death by algorithms. These algorithms will be written, at least initially, by humans, so how we think about consequences becomes vital.
Robot AI could be an important component in many of society’s needs, like difficult labor or risky pursuits, like space exploration. Currently, we have a rover on Mars and if we are to take practical steps towards a manned mission we will most likely need droids to arrive first, testing logistics, insuring the capacity to fuel a return flight etc. In order for this to work, we need robots and the more they resemble C3PO, i.e. autonomous, the more helpful they can be on a mission.
However, Elon Musk has been outspoken about his concerns, suggesting that unchecked AI could be summoning the demon and that AI is humankind’s biggest threat. His comments might seem difficult to reconcile considering Musk’s continual investment in AI companies and specifically his recent launch of a billion dollar research company, OpenAI. However, he actually created this venture with the idea that we must be cautious and diligent with AI.
Musk, along with 20,000 cosigners, including Stephen Hawking, Noam Chomsky and Steve Wozniak to name a few, signed petition to ban autonomous killer robots. This may seem unnecessary and like extreme paranoia, but when you consider that all inventions seem to become weaponized, if there original design isn’t as a weapon, then it seems like a reasonable concern.
The petition is basically that robots must be attached to human control before they can be unleashed as killers. So technically, you could still die at the hands of a robot, as we see in current drone warfare, but not autonomously. This subject is also not some far-off into the future concern, either. Most agree that the technology required for autonomous robots is only a few years away, not decades.
On the utility side, we will soon see autonomous cars, and service robots. But pay extra care to the details of your living will when robot nurses are charge of the plug.