The United Nations recently voted to begin formal discussions on autonomous weapon systems, with 116 of the world’s leading robotics and artificial intelligence experts responding by calling on governments to simply ban them.
The coalition, fronted by Tesla’s Elon Musk and Alphabet’s Mustafa Suleyman, claims this is a dark road the world doesn’t want to go down. Aimed at the International Joint Conference on Artificial Intelligence, a letter from the group warned the U.N. not to usher in the “third revolution in warfare” (following gunpowder and nuclear arms).
While I’m not about to suggest there aren’t serious risks involved with weaponizing thinking machines, it does seem lightly hypocritical for Musk to condemn them over a lack of trust while continuing to champion self-driving cars. Apparently, technology experts feel a Terminator scenario is thoroughly unacceptable but a potential Maximum Overdrive situation is just fine.
To be fair, the issue isn’t so much about sentient and vengeful machines but the unpredictability (and potential mishandling) of autonomous technologies. “Once developed, lethal autonomous weapons will permit armed conflict to be fought at a scale greater than ever, and at timescales faster than humans can comprehend,” read the coalition’s letter. “These can be weapons of terror, weapons that despots and terrorists use against innocent populations, and weapons hacked to behave in undesirable ways.”
As modern cars have already been proven to be hackable, autonomous vehicles will be subjected to even more risk than their present-day counterparts — which isn’t a particularly reassuring concept. Musk has been vocal on the inherent dangers of A.I. for some time, which he still intends to use in his future self-driving Teslas.
It’s more than a little strange that over a hundred experts in autonomous machines and artificial intelligence convened to warn the globe about the technologies they are currently developing. But I suppose J. Robert Oppenheimer more or less did the same thing — and the world probably should have listened.
This is the second time scientists have come together to denounce lethal autonomous weapons systems. Two years ago, thousands of A.I. researchers signed a petition calling for a global ban of the technology, helping push the U.N. to enact formal talks on the subject.
However, according to the The Guardian, some of these weapons already exist. Samsung’s prototype SGR-A1 autonomous sentry gun is rumored to be deployed on the South Korean border, capable of threat assessment and engagement decisions. The United Kingdom’s Taranis drone, in development by BAE Systems, is also believed to possess autonomous functionality. But they’re not the only examples. The United States already has an experimental self-driving warship capable of hunting enemy submarines without human control, and both the U.S. and Russia are working on unmanned ground vehicles that range in size from a shoebox to a fully-fledged battle tank.
Ryan Gariepy, founder of Clearpath Robotics, said “unlike other potential manifestations of AI which still remain in the realm of science fiction, autonomous weapons systems are on the cusp of development right now and have a very real potential to cause significant harm to innocent people along with global instability.”
They actually sound a little further along than self-driving cars — which, again, nobody is complaining about, despite using similar systems to make the decisions your safety hinges upon.
Obviously, I’m being hyper critical and hyperbolic. Even though autonomous vehicles can be weaponized, they aren’t directly comparable to legitimate weaponry. However, it has been a free-for-all as to how this technology is to be implemented, with less oversight than you’d expect. Nobody seems to know how best to do it, yet every major automaker is trying to get it to market as fast as possible. The technology doesn’t need Dr. Ian Malcolm’s reality check of, “Your scientists were so preoccupied with whether or not they could, they didn’t stop to think if they should.” But it should, at least, be in the back of everybody’s mind.
Unlike computer-controlled weaponry and bringing back the dinosaurs, self-driving cars will save lives. There are too many lousy drivers on the road for them not to. However, it would be nice seeing the people behind the technology expressing the same kind of care and seriousness for autonomous automobiles as they do for drones.
[Image: Ford Motor Co.]
Death Race 2017.
And, photo caption:
“TARGET ACQUIRED”
Your comparison between self driving cars and automated weapons is invalid in so many obvious ways. While TTAC like other websites try to be provocative to generate clicks (this comment is your 1 victory) at least find some grounds in your argument to not lose all respect.
Obvious reasons that this comparison is invalid
1. Intent of weaponized systems is to kill as efficiently and reliably as possible, vs. an auto where the intent is to drive safely
2. With weaponized system you have potential to trigger a war and cause a level of damage that will commit world powers to a war of annihilation of humanity, much more than the rate of accidents for self-driving cars which is comparable to haman piloted ones.
To be fair to the author he did say he was being hyper critical.
Nothing to stop the AIU developing and then cars communicating and deciding to kill their occupants or pedestrians – science fiction.
This. Thank you.
The Model S killed a dude pretty efficiently and reliably, even though it’s intent was to drive safely. And just wait until we start trusting AI to decide who to kill and who to save.
Musk, Suleyman and the rest of their ilk, needs to get back to getting busy going to Mars and other hard to get to places. Stop wanking around with the kind of useless rabble, whose only contribution to anything is running around trying to dictate and ban what their betters can and should be doing.
Weapons have been increasingly autonomated for decades; heat seeking missiles being the most obvious example. All such weapons do, in fact all any machine does, is allow people to make decisions higher up the abstraction tree. Instead of the human having to make detailed decisions about a missile’s direction when fired, he can “instruct” it to hit this and that heat emitting source, and leave the automation to make the detailed decisions about how to do that against an unpredictably moving target.
Ditto for anti air, anti missile and even sniper systems simplifying targeting of moving targets. Next up is sniper rifles that attempts to determine the optimal instant to trip a trigger, based on high value target facial or other features. Which will, over time as recognition gets more reliable and proven, allow for fairly autonomous unmanned snipers to be inserted where risk to personell would be too great.
Since there never was, never is, nor ever will be, some sort of hard line between what is autonomous and what is not (heck, a landmine makes autonomous explode or not decisions in even the most primitive of theaters every day… Imbuing those things with enough brains and sensors to chill out when the guy stepping on them is a 4 year old, while exploding if he is a military dude, would be nothing but a boon to both kids and the guys who feel compelled to spread those things around so liberally….), increased autonomity will keep creeping in, regardless of what a bunch of busybodies with preferential media access happens to opine on the matter.
Exactly. Very insightful and informative comment.
There are several times as many guns in the US as there are cars, but guns kill a fraction as many citizens. If guns were as lethal as cars, every municipality would have a fire arms fatality rate to rival Democratic controlled dystopias like Chicago, Detroit and Baltimore. The world would be a saner place without Elon Musk.
Also, if we’re talking about automated guns (rather than heavy artillery/ missiles etc), let’s not forget that the commies employed tripwire-triggered automated machine guns to kill their own subjects who tried to escape their country-sized gulag.