January 05, 2005
More helicopters with AI: The Unmanned Combat Armed Rotorcraft (UCAR) is an Army program to create teams of autonomous attack helicopters that work together and with human controllers.
What seems to be somewhat new here is the combination of a combat/attack mission, a high level of autonomy (as opposed to current drones like the Predator), and the focus on teamwork (this reminds me of some bots I wanted to code for a BattleTech mux, with member of a bot team cooperating with each other and with human-controlled mechs via natural language-based communication).
From an article in Flight International:
Left to their own devices,
the UCARs will co-operate in the same way
as manned aircraft, he says. “Manned aircraft
overlap their sensors and weapons,
cover each other and relay communications.
A team of unmanned vehicles brings
the same benefits,” he adds.
According to Northrop Grumman,
autonomy enables engagement options
such as “protect self”, in which the UCAR
will avoid or engage threats to itself; “protect
team”, where it will engage threats to
any UCAR and sacrifice itself to protect the
manned aircraft; and “protect friendlies”,
in which the vehicle will engage threats to
The air mission commander - the copilot/gunner in the front seat of a Boeing
AH-64D Apache in the Phase 2 simulations
- interacts with the unmanned rotorcraft
not as an individual but as a team. “The
UCARs nominate a team lead to act as the
focal point for interaction with the
manned aircraft,” says Woodbury. The
team lead can change as the mission
unfolds, as was demonstrated in Phase 2.
“If an aircraft had to go back, it seamlessly
passed the role to another.”
Other members of the team take the lead
in other functions, including interfacing
with external information systems and
making the data available to the team.
UCARs will take information from the
Multi-sensor Command and Control
Aircraft (MC2A) or the Global Information
Grid. “They can also push information back
into the common operating picture, and
that is powerful,“ says Greg Zwernemann,
Northrop Grumman programme director.
Northrop Grumman's concept involves a
platoon of up to six air vehicles: two scout
UCARs flying at higher altitude to provide
wide-area search and communications
relay; and four attack UCARs flying at lower
altitudes to identify and engage targets.
“The scout UCAR looks ahead using
SAR/MTI, while the attack UCAR uses
EO/IR and ladar to target and execute
using onboard weapons,” Zwernemann
says, adding: “The two scouts have different
look angles and fuse the information
to improve target identification.” The
UCARs are identical and interchangeable.
“If an attack vehicle runs out of weapons,
it can change roles with a scout vehicle.”
I think this is amazing stuff. And inevitable. The main difficulty probably won't be in designing and implementing an architecture for this level of autonomy (even if it's not perfect, it will be adequate) but in coming up with sensor technologies and strategies for interpreting sensor inputs. You need enough confidence in your sensing that you're willing to let these things to decide, on their own, to attempt to kill something. And they, just like humans do now, will make mistakes and kill the wrong people. I'm sure the engineers writing the software are thinking about that possibility.
By the way, I'm totally conflicted here. Working on this sort of thing would be incredibly challenging and fun. But even though I don't think all military work is unethical, the morality here is still awfully complicated.
Update Jan. 5 2005: UCAR, which would have been “the most autonomous unmanned aircraft ever built”, will not be built (anytime soon, at least). And it sounds like maybe DARPA isn't going to be as involved in the development of unmanned military systems as it has been.
Posted by jjwiseman at January 05, 2005 01:45 AM
It appears inevitable because the world's only superpower (I call it "home") is hellbent on an artificial arms race.
The Architect of the Cold War George Kennan pointed out in his NYT article, "The GOP Won the Cold War? Ridiculous," that we did nothing but idiotic harm with our extreme militarist posture. "Nobody--no country, no party, no person--`won' the cold war. It was a long and costly political rivalry, fueled on both sides by unreal and exaggerated estimates of the intentions and strength of the other party."
Further, bin Laden of all people pointed out right before the election that both he and Bush share the goal of bankrupting our national treasury. Bush's economic incentive is providing welfare straight to industry. On idiotically expensive and useless tech which doesn't build useful capital -- it just kills those who politicians dub enemies.
When I learned about the actual forces behind tech, to explain why Lisp bombed, it became clear that tech going in this direction isn't inevitable. We choose.
Which btech mux were you on and when?
HEY! This is fantastic! Will the system feature enough AI to determine the malicious intent of the teenage kid who's forced to venture out into the street to find clean water because the rest of the UCAR team killed her caregivers and destroyed her city's water supply? That would be keen!
What the fuck is wrong with you? Just look at that fucking picture you posted. Jesus christ. Do you know who lived in those buildings that are now serving as traction for the tank? PEOPLE, that's who, and almost certainly people that had no reason to have their buildings demolished. How could you possibly have any ambiguity about wanting to develop or promote such a future?
I can't fucking understand it, but there seems to be some unbreakable rule for much of the Lisp/AI community to show unwavering deference to DARPA and unquestioning enthusiasm for any weapons system that even mentions AI.
God damnit... Can't you do anything better?
Like I said, it's complicated. And a statement like "I don't think all military work is unethical" is pretty far from "unwavering deference to DARPA" and "unquestioning enthusiasm".
A related project is MICA: Mixed Initiative Control of Automata-teams.
Not in Lisp, unfortunately.
The last sentence of the article you mentionned is pretty telling though:
"..Darpa tells Aerospace Daily it remains "committed" to building unmanned systems. But, when it comes to armed robots, the Pentagon higher-ups may not be committed to Darpa..."
There is a very large wiggle room for DARPA to continue developing unmanned systems for support mission only. I am pretty sure that even though the internet was developed by DARPA, the pentagon top brass never saw it as their unique way for coping with nuclear attacks (MilStar,...)
Didn't you see THE TERMINATOR movies?
I hate to get involved in responding to a rant made by someone who feels so strongly about his/her argument that s/he had to post it unsigned, but I find I can't resist in this case. *g*
Actually, you know something? *I* can't tell from that picture if the "UCAR team" is the reason that building is in ruins, or if they are occupying position that had been reduced to rubble by someone else's forces. I can't tell if the UCAR team is the aggressor or the defender. I can't tell if the UCAR team is fighting a hold actioning to allow the teenage child, her caregivers and other "innocents" to safely evacuate in the face of otherwise certain annihilation, or if the UCAR team is the cause of their alleged destruction. All I can tell for sure is that it's a rendering of a static scene. A static *FICTIONAL* scene.
I understand your underlying point, Unsigned Poster. Do you get mine?
In case not, let me spell it out. As a strong supporter of having well equipped military AND as very vocal, very early critic of the current Bush administration's foreign policy, military policy...heck...pretty much every policy of this administration, I can tell you that I want our soldiers to have the best weapon and support systems possible, and the best leadership to know when and how to deploy said systems.
The issue isn't that resources are being expended in the development of military related technology. I would argue, the issue is to ensure that those resources are allocated properly, and that military force is used appropriately. There are many issues to be considered in the development of any weapon systems. Moral. Ethical. And yes, even logistical. The fact that these particular weapon systems are "AI-based", I'd argue that those issues are even more vexing. But rants like yours don't advance the discussion one iota, but in fact detract from any meaningful discourse.
The unsigned post that's lighting up this comments section was "hot, "anonymous", and from my end of the scope, right on the money. I don't think all military work is unethical, but the truth is that's practically an empty statement. The ethos, whatever it is, and everything has one, is, finally, determined not simply by intent but also by practice. The ethos of your intent and your work doesn't end with your given assignment in that work, and in reality you don't have any say or any control over how any of it is deployed and employed. What you get is a paycheck, some professional kudos, and maybe, if it's daylight work, some bragging rights (if you then feel 'ethical' about what's done).
What you CAN do and I believe, ought always to do (because it is in your power to, and that being so you can't say you didn't know) is to look at the behaviour and active values (as in played out in practice, not proclaimed or echoed on parade) of the organisation your efforts will be a part of and entrusted to. Here your "sight" is a matter of the heart and NOT the head. And we all know how hard that is when your 'head' is arguing dollars, family and good consumer stuff.
Look, don't theorize about this shit, just take a long, long look at Iraq - Fallujah, Samsara, Najif , and others. There's the truth behind all of your arguments laid out in exacting, deadly, high-tech motion. What's there and made both possible and so by the folks so apparently uncritically admired by some here, makes the UCAR picture look like some artists cartoon-idealized view of what it's like to actually go out and kill some 50,000+ (considered conservative) women and children (yes, not counting men) civilians (Iraqi) in the last 18 months. Now that's not a STATIC, FICTIONAL number.
Hey! Great weapons systems though!, especially the one's Rumsfeld keeps bragging about as being sooo accurate and therefore (his words, not mine), in keeping with an ethical military approach and all. Smell the roses I say, true, working for or with DARPA IS sexy, exciting, and you can do some of the baddest tek stuff on God's green (still) Earth; but that doesn't mean you can't connect the dots between what gets done there and, through the increasing politicization we've seen of the Pentagon and US military missions, what the Pols order in our names and do with our tax dollars (which when discovered we are outraged, apalled and figure outing a "few bad apples" will sort it out)?!.
There's nothing complicated OR complex about the decisions to work on this shit, and a lot of it is so bloody obvious you have to work hard NOT to see the only thing it's designed for. And despite the comparative trickle of stuff into positive civilian applications, the vast majority of this technical ingenuity (and ingenious much of it is) is wrapped down tighter than a butterfly's tit, is deep black, and NO leader in the last thirty years has evidenced ANY public intent to take this bad shit and do good stuff with it (nor would he/she last if they did as the boys do not like people messin' with their toys).
Look up at the missions and event horizons of the organisations and parties you're talking about - are they making you feel safer and are you becoming richer as an American?, and if not, why not? Are bigger, better, and badder weapons/systems going to actually win hearts and minds, and therefore your increased security and prosperity? Have they ever? We've spent some 151 billion bucks since the Iraqi war started and we're just getting started, unless we effectively cut and run. Within that budget we're building 14 huge bases all over Iraq, which means we're not exactly thinking about leaving tomorrow but, namnesia aside, we're not supposed to think about that. But just for comparison's sake, some 8 billion bucks would have paid for the primary school education of EVERY kid on the planet for one generation. Now that's just a few hearts and minds without firing a single bullet.
With respect to rcs's comment: "In case not, let me spell it out. As a strong supporter of having well equipped military AND as very vocal, very early critic of the current Bush administration's foreign policy, military policy...heck...pretty much every policy of this administration, I can tell you that I want our soldiers to have the best weapon and support systems possible, and the best leadership to know when and how to deploy said systems." While I admire your spirit rcs, you do seem to be ignoring in very practical terms the ground truths of the last four years in respect of the political, civilian and military leadership into whose hands these weapon and support systems have actually been thrust. You don't have to guess or hope, look at what's actually going down right now. The "leadership" we have is the one we've got, not some hypothetical "best" somewhere. This is it.
The reality crawling stubbornly and painfully into public view is that "the best leadership" you talk about is busy twisting statutes to permit and enhance torture; expressing surprise, ignorance and disappointment about a "few bad apples" that have, somehow mysteriously, crawled out of a virtuous virgin's command of leaders; and that even after "owning" the country for three years 'the leadership', not any of it, can find even one item (WMD's) of that which they were assuring us was absolutely located (no doubts about it sir!) only where they knew, and based on which, we should give them free reign over our fathers, mothers, sons and daughters (some 1500 gone, some 40,000 grievously injured). Come on, your comment is idealistic at best and certainly naive (nothing personal intended) with respect to these folks.
You're also dead wrong on both sides of your statement when you say: "The issue isn't that resources are being expended in the development of military related technology. I would argue, the issue is to ensure that those resources are allocated properly, and that military force is used appropriately.". The key thing is who makes any of those decisions? And what practical effect do you or I, as citizens, have over those decisions which spend OUR common treasuries and allocate those systems and their uses to ends we would have no truck with? Today's answer is, in pragmaticl terms is f**k all, or, more politely, very little, if any. A lot of folks used to think elections were indeed "an accountability moment", but after the two Floridas, Ohio, New Mexico and a few others, is that still true? And after "special interests" have had their day on the Hill (Washington D.C.), with all due respect, who's left to buy who can "talk sense" to those already bought? Like Josef Stalin said prophetically and knowingly: "Those who cast the votes decide nothing. Those who count the votes decide everything."
"Our job is to give people not what they want, but what we decide they ought to have." - Richard Salant, Former President of CBS News
Re: "We've spent some 151 billion bucks since the Iraqi war started and we're just getting started" ... wouldn't it have been better just to fly over Iraq and carpet-bomb them with real dollar bills (okay, tens and twenties)? Just think about it - no loss of life, stimulate the economy, give the little guy a chance at real purchasing power... Maybe we could try it on a smaller scale in Afghanistan?