Gadgetory

you send a robot into a hostile environment where decisions have to be made quickly and the algorithm will have to take care of that the question is will there be humans in the loop or will they be on the loop or will they be out of the loop killer robots are weapons that would make drones look primitive at least with the drone there is a human being who looks at a computer screen sees the target and pushes the buttons to fire the missiles and kill as we begin to approach the possibility of having machines select and engaged targets we want to be very careful not to cross that line without high-level policy review as technology races ahead as we achieve these fantastical advances what decisions are we going to be comfortable delegating to machines and what kinds of decisions are we going to insist on reserving for the exercise of human judgment we're a long way off from the day when armies of robot soldiers will march in perfect formation shooting lasers across enemy lines but the US military has stated its intent to use robots on the battlefield where they can theoretically carry supplies bring cameras in to dangerous places and yes even kill some experts are calling robotics the new arms race for that reason an international group of scientists professors and activists including the Human Rights Watch and the International Committee for robot arms control are calling for a debate on the questions that inevitably arise with robots that can be used in war i'm martha brood and i'm a member of the international committee for robot arms control I'm a physicist by training and I proposed a ban on autonomous weapons as early as 1988 we go to war and when we see what's going on we decide there's a point we don't wanna go beyond that point if you look at the history of the Cold War there are many incidents where people interrupted the chain of events if you look at the crisis decisions at the highest level that were made during the Cuban Missile Crisis or other major international crises it's always a point where somebody says no don't go don't take the next step and if we automate everything is not going to be that human intervention somebody always has to do that no more otherwise wars will never rent either you won or you lost or it's just too much blood and it's not worth it anymore at some point people say stop but if we outsourced war if we outsource the process of conflict if we make that all a matter of machine decision and we're not going to have that that intervention of the human heart this is going to be just the program that's running good morning everybody for all of us in the DARPA community there's no place we'd rather be even right here today and the reason for that is that a DARPA our mission is about breakthrough technologies for the future to help make our nation in our world a safer a more secure place DARPA is a 55 year old agency in the defense department we were started in the wake of Sputnik it was a real wake-up call for the United States it was a huge surprise and then as now we understood that technology is a cornerstone of our national security people were pretty clear that we did not want to go through that kind of surprise again so DARPA was created specifically to live outside of the rest of the way that we do our science and technology investments to be a projects agency with a specific mission on breakthrough technologies for national security the general desire is to build machines that amplify the effectiveness of people during the Iraq war there was a very difficult problem that the Defense Department faced with improvised explosive devices and so we had funded a number of robots before and DARPA helped to see that those machines could be adapted for getting rid of IEDs really when you think about what robots are capable of today it's very very very early and here at the DARPA Robotics Challenge I think that really comes home to you when you see robots taking you know 30 minutes and not being able even in that much time being able to do the things that we as humans would find very very simple to do but focusing on disaster relief in the context of this challenge I think really allows us to to push the technology first to see what we're able to do today and then start pushing it to the next level so that eventually we will have robots that have far more capable features for disaster relief and I think you know I think for many other applications as well we're not building weapon systems here we're building the underlying technologies that can be used in many many applications and our main focus here is recovery and disaster relief but just like DARPA worked on the internet GPS fiber optics all of which we use every day and so does the military robotics will be the same we'll use robotics in our homes in our work and some of that will be military but not all of them to be sure if somebody came up to me and asked me to build a robot they could fire a gun I'd probably say no I like what I do now I like the research because the research is fun the technology itself doesn't say what it's going to be used for in certain cases of course like a particular weapon and a bullet and things like that those have primary uses for military systems and are not dual use but these robotic systems are very general purpose and you really need to understand that they are neutral with regard to the concern of are they military or non-military so whether the Defense Department funds them or some other company funds them for whatever use that development of technology is in fact possible to end up in any kind of system given the outrage over the military using unmanned drones and lethal strikes many people may not be comfortable with the idea of robots as weapons and most people in the robotics community seem to agree that now is the time to have that conversation before we have killer robots not after million comes down to how people use technology it's a matter of human wisdom and being thoughtful about how technology gets applied I'm very confident that we will have the wisdom to use robotics for mankind's benefit but we can't assume that will happen we have to make that happen this is an issue that well when I first started talking about 25 years ago and people would just kind of stare at me and then ten years ago they would just say oh yeah terminator yeah so ladies you know it was a big joke just in the last five years it's you know the giggles of stalked people are realizing it's too serious you know our job at DARPA is to invest in advanced technologies and we pursue them because of their promise but robotics is a great example of an area where we also recognize that that in driving those technologies forward we're also raising a whole host of very important broader societal questions I don't have a position in terms of whether the US should or should not sign any kind of and I think that the directive that the DoD has itself signed with regard to legal autonomy has been it's really good it's been very carefully thought through it says that the primary concern is one of reliability you'd want to make sure that if these systems decide within a particular set of instructions from a human operator whether to go to one place or the other that that choice is always in line with the intent of the operator so it really reflects back to a human beings choice there's someone like me comes along and says we should have a hard red line we should not cross that's what they don't want to hear instead what they want as well we should think about what the ethical rules are which is just a way of deflecting the concern and and and saying we're gonna do it you know we have their rules about how we're gonna do it some people will say oh well you know where do you draw the line and I always say well you draw it somewhere that's the important point and you