 Yeah, I'm talking here today about the role of civil society in international arms control. I'm especially dealing at the moment and already for more than a year with the issue of killer robots, or more technically autonomous weapons systems. I think it's a very interesting issue to look at the role of civil society with a very, very central role at place in even setting the agenda, framing issues, and very fundamentally starting an international debate. I have a political science background as well, so I'm approaching this a little analytically. As we are obviously here not a symposium of artificial intelligence experts, or specifically on the matter, I want to hold it a little abstract, also with a few to be able to generalize on our panel discussion. So my attempt in the paper is to understand and explain why these civil society in the field of killer robots autonomous weapons systems is as successful as it is. In order to do that, let me give you a very brief overview timeline of what happened. Autonomous weapons systems in general are not defined at the moment because that's international consensus that do not exist at the moment. So it's a preventive campaign if you want to say so. We understand under autonomous weapons systems militarized robots that can basically take every form. They can be a refitted tank or refitted jeep or a swarm of vehicles that fly comparable to today's toy drones. The decisive criteria here is that they have a certain degree of autonomy from human control that is especially in the critical function of selecting and engaging human targets. So that's the technical part. The international debate if you want to say so started between 2004-2006 on a very, very technical level. It was especially Professor Arken of the Georgia Institute of Technology who started to theorize about the potential of including moral agency and autonomy algorithms. He worked from 2006 on with the US military to determine the potential of a military usage of autonomous robots. That was very much confined debate to artificial intelligence experts at the time but sparked among the academic experts opposition movements. And it actually started here in Great Britain. Noel Shacky published in 2007 in The Guardian an article. He is a roboticist by profession. He published an article in The Guardian warning against the dangers of autonomous weapons systems and proposed strong international regulation. This was the decisive step in this very first phase because through publishing in newspaper he reached a kind of mainstream. For that it were all artificial intelligence journals which of course have a very narrow reach. Through The Guardian he reached another kind of community. He also reached arms control campaigns. And afterwards in 2008 I believe for example the Landmines Action Campaign as the name says a campaign that dealt with ban on landmines joined in a call to ban or preventively ban autonomous weapons systems. We have here a very interesting factor that I identified and that's my ultimate goal in this phase of my PhD research to identify factors for successful campaigning here. That we have the benefit in the arms control sector of pre-existing networks. We have activist networks like the Klusum Nations Campaign, like landmines actors who really know how to do it. They have the knowledge there, the personnel, they have done it in other issues before. They know the when used to the government, they know the international platforms and a lot of experience in the activist field here. That's a real asset. So when landmines action calls for something like that it is of course listened to in its network. You activate a lot of other actors as soon as somebody takes up this issue. And that's indeed what happened in the arms control community in the international arms control network if you want to tell it this way. Increasingly more actors, of course there is still a very insider discourse but increasingly more actors become aware of the issue. Take it on. And in 2010 we had the first transmission, if you want to say so, to the international stage by a report of the UN Special Rapporteur on extrajudicial killing who mentioned lethal autonomous weapons as a problem on which the international community has to look in. So with this introduction to the United Nations agenda where states had to deliberate on that, take positions you suddenly of course attracted also the attention of the big players, I want to say so. You have Pax Christi or Pax joining on the issue, publishing a report on autonomous weapons systems in 2011. You have in 2012 the British NGO Article 36 very influential arms control NGO here and very big success considering that it all started with roboticists in 2012 Human Rights Watch published a landmark report called Losing Humanity and called internationally to be aware of the legal problems and the ASIC problems involved in autonomous weapons systems. As soon as you have Human Rights Watch as a really major player you can attract a lot of journalist attention. It is not surprising that they should gain some traction here. And indeed in October 2012 Human Rights Watch published this report. November 2012 the US Department of Defense issues directive 3,000.09 in which it effectively, as you cannot say but it puts a moratorium on the implementation of autonomous weapons systems into the US services. That means it gives the first definition what do we understand under autonomous weapons and it says as we are not sure at the moment so what to do we want to remain at a level of human control that is acceptable at the moment. So they for the moment postpone autonomous weapons systems fully autonomous weapons systems under 2020 or something and have thus opened first of all a big policy window of course in the United States but also internationally. We are now at the moment where a lot of flux and theoretically something could change because we are at the very beginning of a process at the moment you wouldn't have a lot of losers if there would come some kind of international regulation or ban. And yeah it's very important that they reach this policy directive already to not create any sunk costs in an issue. 2013 then a couple of months later in London, in Britain, a network of NGOs launched a campaign to stop killer robots Human Rights Watch, Article 36 and a lot of other NGOs from all around the world joined this platform to funnel their resources and coordinate their advocacy and not the least of which also gained some social media attraction and publicity because until then it was often called very technical, lethal autonomous weapons systems or something like that and that's of course not very handy for headlines while when you suddenly can talk about killer robots I bet that has some journalistic attraction. So yeah it's a major framing success when you have a lot of actors in the network starting to talk in terms of killer robots because that's an easy term that will probably transmit through the institutions and it also causes and that's also what I say immediate instinctive emotional reaction I want to say if you hear about killer robots most people would instinctively say we don't want something like that. That's general population that's also within the institutions of the government and very important here is that through these actions the campaign succeeded to change the type of the issue What in the very beginning as I said a very technical issue between artificial intelligence experts that wasn't very attractive to mouse people Now they were talking about killer robots and about ethics that's something where you can have instinctive reaction to it's an easy issue we say and that's easier to sell and that's easier to lead people to their attention and reaction that the campaign would like them to have Very important here also the celebrity support We have a lot of important and well connected people joining the call Jody Williams from the Endmind campaign Nobel Peace Prize Laureate, a lot of other laureates joining of course here Stephen Hawking signed a public letter Elon Musk and every time people like that do say something journalists are going to write about it it has really some supportive function and I fear some data we have unfortunately of course since it's not really an existing issue at the moment very limited data but we have this one carpenter survey and conduct is a little research on the public opinion in the United States on these autonomous weapons We see we have in the general population a clear majority that says even so you have to say people were asked like that we received a little explanation but there are certainly no experts but they say just instinctively that's alright we do not want that even more importantly from my point here you have in the very institutions that would be in charge of employing these machines even stronger distrust of these autonomous weapons systems that's surveyed by military status we can see here people who are currently active in the military or serve previously are distinctively more opposed to autonomous weapons systems than the general population that means the very people in charge of doing that are very skeptical about that and that's certainly also benefit for the campaign that you have a lot of supporting forces also within the institutions same is true for political active people this graph shows the stronger your political alignment the more liberal or the more conservative you are the more likely you are to be against against the graphic is unfortunately a little not formatted but anyway it tells you the more politically convinced you are the more you are against autonomous weapons system and that's of course important to the political side because the more politically convinced you are the more likely you are to vote so voters are not convinced of the issue another factor as I already said you have a lot of connections in place I made here to visualize the network what we are talking of we have in red the killer robot network if you want to call it this way and the graphic shows the website connections among the actors it's one way to represent a knowledge network here and we say it has a lot of connections to the cluster munitions network and to other arms control networks and they all work together and are linked to the United Nations and that's a very interesting thing I will close with United Nations which provide the framework for a lot of talks at the moment that are going on in the state side within the convention for conventional weapons systems they provide it as information sources for the most time until very recently mainly sources that are connected to some kind of campaign platforms so the campaign the civil society was able to be the prime information provider for all the state actors that are now starting to deal with the issue simply because everything else is very technical, artificial intelligence information that's not very accessible to governmental decision makers of course so by taking this kind of information monopoly they are certainly very able and successful in determining and framing the issue of development and that's only a little overview what we have at the moment currently 22 countries support the ban not the most important ones of course not surprisingly but the point here is that none of these countries in the killer robot says we want to have them or we are actively opposed to ban the argument at the moment is we don't need it because they do not exist so they have already managed to frame it in a way that nobody really wants to be against them that's certainly a success thank you