Let me start out with a quote from a New York Times article on Ethical Killing Machines:

… a 2006 survey by the surgeon general of the Army, which found that fewer than half of soldiers and marines serving in Iraq said that noncombatants should be treated with dignity and respect, and 17 percent said all civilians should be treated as insurgents. More than one-third said torture was acceptable under some conditions, and fewer than half said they would report a colleague for unethical battlefield behavior.

Those statistics are… disheartening to say the least. While a majority of soldiers are good people I’m sure, the bad apples are the ones that get the bad press (Gitmo, anyone?). The ethics of our soldiers are certainly not under the control of anyone but the soldiers themselves. That’s pretty worrisome.

One solution, of course, are robots. As the military moves closer towards autonomous robots that do all the fighting for us, some are asking the question of whether we can build ethics into robots (which is the topic of the New York Times article quoted from above. Anyone who has read “I, Robot” and is familiar with the three laws of robotics may get worried about the idea of robots with ethics.

Some may frown at the flexible ethics of some soldiers and demand robots who cannot break their ethics codes. I think a greater number of people will be worried about robots with their own code of ethics, and would rather not see robots given such autonomy at all. However, this is a question that must be asked. It’s only logical that we end up building robots that can fire on enemy soldiers, tanks, or buildings. The order to fire can be left up to a soldier at some distant command center, but those decisions take time and someone who is not directly in the situation may not have full situational awareness. It may be a better solution to let the robot make the decision.

Anyone who has taken an ethics class knows that there are no black and white areas, and anyone who has taken a computer science class knows that computers are bad at gray areas. This seems to be an issue of Levels of Automation, which is a huge area of research aimed at determining how much responsibility to give to a human and how much to a robot.

If the robot does everything and a human just has to press Yes all the time, the job gets boring and the human doesn’t pay attention. However, if the human has to do a majority of the work, why bother with a robot at all? This is a vast field of research, with an abundance of experimentation and case studies to draw from. Thankfully, while some roboticists and ethicists work on installing an ethics module into automated robots, most people will require some human in the loop and somehow responsible for who lives and dies. This requires the knowledge of a human factors specialist, and it is comforting to know that ergonomists will be on the forefront of determining how soldiers interact with these automated killing machines.