Fox News: Military Robots Will Kill Everyone [Reckless Fear-Mongering]

Writing by Evan Ackerman on Friday, 27 of February , 2009 at 3:41 am

Experts Warn of ‘Terminator’-Style Military-Robot Rebellion:

Autonomous military robots that will fight future wars must be programmed to live by a strict warrior code, or the world risks untold atrocities at their steely hands.

The above headline and quote is how Fox News decided to lead their story on an Office of Naval Research report on military robot ethics. You can read the full report here (*.PDF).

Firstly, just let me say that this is certainly a relevant issue, and it’s something that needs to be carefully thought about, discussed, and addressed. But this is not how to go about doing that. The media has tenaciously latched onto the idea of military robots killing innocent people, because it happens in the movies and it gets people upset. But it’s not accomplishing anything, and is just encouraging the public to be that much more resistant to (and even afraid of) robotics. This sucks, and it sucks that so-called “robotics experts” are encouraging this sort of behavior, since it sets us all back.

I’d like to spend a little time discussing this issue, mostly because I think it’s important, but partially because I’m seriously pissed off. Don’t worry, I’m not going to rant (much), but I am going to try to explain why I think battlefield robotics are the future and why it’s possible and reasonable and a good thing, and also why in a lot of ways, we are experiencing (and benefiting from) that future already.

Perhaps the most relevant issue here, although it may not seem like it, is what we decide to call a robot. By many definitions, we’ve been using fully autonomous robot weapons systems for years now, in the form of missiles. What I’m thinking of specifically are fire and forget anti-radiation missiles with loitering capability (example)… These missiles are fired at a radar source, and if the radar shuts off, they’ll wait around until it starts emitting again, reacquire, and destroy it. In this case, after the weapon is launched, it is autonomously locating and destroying a target which may involve human casualties. Strictly, how is this different from (say) a fully autonomous drone that’s sent on a mission to destroy a radar source?

I think the difference is that we don’t generally think of missiles as robots. But in a lot of ways, they’re more robotic than something like MAARS since once launched, missiles depend on their own artificial intelligence to guide them to their targets. Currently, as far as I know, there are no fully autonomous armed mobile ground robots that have this capability: the robots are not directly in control of their weapons systems, and the ultimate responsibility to fire the weapon is under the control of the human operator.

Part of the reason that people are getting upset is that eventually, these robots are going to get control over their weapons. Not any time soon, but eventually. And the worry is that for whatever reason, they’ll start killing everyone (or more specifically, civilians in war zones). In the ONR report, they discuss the problem of a bunch of modular code written by different people exhibiting unexpected and potentially dangerous behaviors. With regards to this, Fox News says:

“No individual could accurately predict how the various portions of large programs would interact without extensive testing in the field , an option that may either be unavailable or deliberately sidestepped by the designers of fighting robots.”

I’m not even going to bother responding to that, because it’s not “news,” it’s not designed to communicate information, it’s designed to scare people. And that’s shameful.

What I am going to respond to is the idea of extensive and complicated testing, because this, and not extensive and complicated initial programming, is how military robots can be made safe and reliable. Run them through all kinds of scenarios, try and get them to make mistakes. And when they do, figure out what went wrong and fix it (a task which in many cases is relatively easy with a robot, compared to a human) until they don’t make mistakes anymore. This is how we train humans, isn’t it? And most of the time it’s effective… Training and experience helps humans to develop battlefield ethics and distinguish (for example) civilians from enemy combatants.

Critics of battlefield robotics seem to neglect two important points. Firstly, the whole point of arming these robots is that they can take the place of human soldiers. They won’t be as capable or as effective, certainly, but they’re replaceable and they don’t have families. This, more than anything, is why we must pursue these technologies. The second point is that humans, for all of their ethics and morality and experience and training, make all kinds of mistakes. Sometimes, situations occur that are beyond the scope of reasonable preparedness, but nonetheless decisions have to be made and actions have to be taken. The key is to learn from these mistakes to make sure they are not repeated, and this is the same thing that can be done with robots. Accidents will happen. Inevitably, an robot is going to make a mistake and hurt someone. We accept that in war, accidents happen and mistakes are made. It’s awful, but it’s war. Why should this not be the case for robot soldiers as well?

We’re already at the stage where many of our weapon systems are semi-autonomous, but the “semi-” only comes in when a human operator has to push an “okay” button allowing the system to engage live weaponry. This is an adequate safeguard in most cases (or at least, giving a human ultimate responsibility makes us feel better), but it partially defeats the purpose of an autonomous weapons system: the ability to act faster than a human can. And in some instances (such as anti-mortar and anti-RPG systems), the human has been removed from the equation for this very reason. It’s already happening, and these robots are saving lives.

Personally, I believe that semi-autonomous (and eventually autonomous) armed robots are a necessary and positive step towards reducing human casualties in war zones. I also believe that autonomous robots can be, and will be, able to operate safely alongside human soldiers and civilians. It is certainly important to discuss how we can make sure that risks are minimized, but we have to acknowledge that many robots are already armed and already autonomous and already doing their jobs reliably and successfully, and recognize that prophecising the worst and then sensationalizing it damages the robotics community as a whole.

[ Fox News ]

Further discussion on this topic is available over at Danger Room.

Update: I just wanted to draw your attention to a relevant article that we posted last year about uncommanded movements from armed TALON robots. What happened is sort of a microcosm of this whole story… A few robots, while undergoing testing, had some minor problems. These problems were vastly overstated and people freaked out. Meanwhile, the minor problems on the TALONs were fixed (redundantly), but the damage to the bots reputation was done, and people were scared. This is how it’s supposed to work, though… You’re supposed to test the robots, find the problems, and fix them so that they don’t happen again.

Comments (16)

Category: Artificial Intelligence,Military

• Add this post to
• Del.icio.us -
• Digg

Name (required)

Mail (will not be published) (required)

Website

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>