Is a robot that is able to learn and take decisions “responsible“ for its own actions? Or is the company that manufactured the robot liable for the damage that the robot may cause, regardless of whether or not the machine was programmed to hurt anybody? Or is the owner of the robot responsible for its “bad habits“? When a pet robotic dog will accidentally hurt a toddler, who will be responsible? The robot itself, the owner of the robot, or the manufacturer? Moreover, if a corporation is found guilty of fabricating robots that “learn” to harm people, who will go to jail, since corporations are non-human legal entities with “No Soul to Damn, No Body to Kick”?

This scenario is certainly not a contemporary one, but increasingly sophisticated and autonomous systems are being developed in both military and civilian applications. First and foremost, autonomous weapons are being developed to substitute humans in very risky tasks. Robots equipped with machine guns, unmanned fighter jets and tanks are in very high demand. Col. Chris Chambliss, commander of the 432nd Air Expeditionary Wing, says that machines are so useful that “We’re well aware that there’s a virtually insatiable appetite for these systems and their capabilities. We know we have to generate more of them as quickly as we can, and that’s our goal.“ (read the article)

As technology and our ability to replicate perceptual and higher cognitive functions improves, these weapons, as well as more innocuous instantiations of intelligent machines, will progressively be able to learn, make decisions and reason in a way similar to a human being. Robots designed to interact (and assist) children and the elderly are being pioneered in advanced markets such as Japan (see, for instance, the Mitsubishi Wakamaru). Busy people in modern societies often have little time to provide the appropriate level of care for children and elderly, and many companies are heavily investing in creating increasingly sophisticated personal robots able to engage and take care of human beings. If you think this is altruism, think again: the market size for personal and service (e.g., roomba) robotics is estimated to reach almost $8 billion in 2010.

As a researcher, I am well aware of the scientific and technical challenges that lie ahead of us before we can achieve truly intelligent, autonomous systems. But I believe that an equally difficult problem lies ahead of legislators, lawyers, and judges when trying to define, for the first time, the concept of “intention“ in an autonomous, non biological system. As the number of robots and their range of application increases, there will be a parallel increase in robot-related litigation. Robots present a unique challenge for lawmakers: unlike the car parked in your driveway, they are able to learn similarly to biological systems. The ability to modify their behavior to better adapt the environment and the varying need of the owner (e.g., learn to engage the child or take care of the elderly) also poses the problem of unpredictability of the robot’s behavior, and the possibility that this behavior could cause harm.

While waiting for the legal responsibilities of autonomous robots to be sorted out, I would keep your roomba on the leash….