Musk, Woz, Hawking, and Robotics/AI Experts Urge Ban On Autonomous Weapons

An anonymous reader writes: An open letter published by the Future of Life Institute urges governments to ban offensive autonomous weaponry. The letter is signed by high profile leaders in the science community and tech industry, such as Elon Musk, Stephen Hawking, Steve Wozniak, Noam Chomsky, and Frank Wilczek. It's also signed — more importantly — by literally hundreds of expert researchers in robotics and AI. They say, "The key question for humanity today is whether to start a global AI arms race or to prevent it from starting. If any major military power pushes ahead with AI weapon development, a global arms race is virtually inevitable, and the endpoint of this technological trajectory is obvious: autonomous weapons will become the Kalashnikovs of tomorrow. Unlike nuclear weapons, they require no costly or hard-to-obtain raw materials, so they will become ubiquitous and cheap for all significant military powers to mass-produce."

Comments

Re:Futile

[Anonymous Coward]

Start with land mines. These are autonomous weapons with little or no AI and have done far more devastation to civilian populations.

The AI arms race will basically be to develop more accurate enemy identification. The "low" tech AI will have more "friendly fire" incidents than the "high" tech AI.

The old game of nethack warned not to genocide shopkeepers. If you genocide them you would kill all humans, including your own character. This seems applicable to AI autonomous weapons.

On the other hand, a theoretically competent AI running a weapon could make choices of not engaging despite a command or even an attack on it because of the risk of civilian or collateral damage. A human holding a weapon would not necessarily be able to make the dispassionate trade of self-sacrifice for some number of strangers or monuments.

Re:Drones

[DarkOx]

Admittedly have never been an infantry man, pilot or any other sort of military man myself I still suspect its much easy for a guy sitting safely in chair to make a moral decision about a target, than it is for a guy in a life threatening situation to do so.

A drone operator can loiter around a target for a long time until he or she is confident said target is properly identified. A jock in a fighter-bomber does not have that luxury and also exists in constant fear someone is going to pop up with an anti air craft device, that will end his life. The drone operator has to worry an anti air craft device will ruin his afternoon with extra paper work. I known which one I'd rather imagine hovering over me deciding if I an enemy combatant or just a guy going out to milk the goats.

The separate question is does done warfare lower the barrier to entry such that conduct operations in theaters that would forgo if it meant having the infrastructure and associated costs of supporting large numbers of manned air craft in the area. This is over great concern. If we make warfare to easy we might find ourselves doing more of it. I am not buying the argument though that drones are equivalent to mindless kill bots or worse than the existing maned alternatives in any given situation all else being equal.

 

Re:Futile

[MightyMartian]

It's similar to the situation at the end of WWI. Versailles called for wide-ranging disarmament among all the belligerents, which was all well and good in theory. In reality, of course, a great deal of the R&D that had gone into new weaponry; tanks, planes, ship designs, and so forth, still existed. In fact, the most valuable commodity of all, the German plans for the 1919 campaign that never was, still sat in archives, just waiting for someone to come along and dust them off.

The cat is out of the bag, has been out of the bag for a few decades now. When most of us look at devices like Mars Rovers, we're impressed by the technology and science, and yet that very same technology is easily adaptable to building autonomous weapons. Even if the Great Powers agreed, you can be darned sure they would still have labs building prototypes, and if the need arose, manufacturing could begin quickly.