 Ie, I do research, I do a huge amounts of data research and in doing that I use lots of computers. The computers we have today, very...we like to make ourselves a victim too much sometimes. We think that these companies and these governments have all the power. Well actually if you can use a computer well you have an awful lot power too and in fact the current restrictions we are seeing gallwch bod yn gwrs ddefnyddio'r cyffwil ffyrdd ymdweud i ni o gyfwyrdiau a'r cyffwilio trwy'r prwyprwynd. Mae'n tra fydda chi'n siarad gwybod i gynhyrch ar y gŵr i gyd, a'n gallu cyffwil i gydig oedd y cyffwil. Yn gallu gwleidio'r hyn a hwn oedd i gynhyrch, mae Unedol fforddau Wythgrif iawn, a that is Croudton that is actually how I got to know Croudton first I walked across it but I can tell you what Croudton is I can tell you what you're looking at but it wouldn't tell me what it does and that's because today the function of technology is completely separate from its form the form and function of technology is changing .. cultural, and it's all about data. Data is the new financial currency. Data is the stuff that whole industries are now being built on. Data is something that you generate but you do not own. When you are getting something for free, you are not the customer. You are the thing that's being sold. And that is a new world we're in. Data is absolutely meaningless, unless you are able to organize it. Gîm ydych chi'n bach â yw metodwyd drwy adnewidyn yn fathau. Mae'r meddwl herefau eich meddwl iswyd wedi ychydig. Ychydig yw bethau mewn ddataethaeth ar ddataethaeth. A ydych chi'n gweithio gweith? Mae hyd y�k gwylleth o ddataeth. Mae wertho ffonaidd, yn fwy o gweithio, yn fwy o fideo, yn fwy o pitches, yn fwy o gweithio gyda gael gael i gwylltaeth. Mae hynny'n gweithio gweithio, mae hynny'n ddim yn digwydd. Yn ddim yn digwydd, the military internally claro. The military internally itiama gwaith arna the military internally itiama gwaith arna the military internally itiama gwaith eich tynnu they have to annotate all the stuff they collect with metadata and data really is, you have what´s called the collection context that´s where it came from who got it and what sensor was used because you have to know that to interpret it but then you have the really dodgy stuff the information context y cyfnod. A chi'n ddiw? Dydw i ddim? Dydw i'n ddim yn meddwl, y rhai sgwrn? Ond byw'n meddwl ydw i'r cyfnod o'r bwysig ddyddiaeth. Felly, gobeithio, diaom mae ddyddiau wedi ddyddai'n ddydd i'r cyfnod i'r dydd. Ond ydych chi oeddwn yn ei ddweud, dyna gweithio'r ddwy'r llwyth. Yn swyddi ddwy'r cyfnod ar y dyfodol iawn i'r ddyddyn nhw ychydig ddweithu hyn. is the reason you get send somebody else a direct mail shot or a person who is dead gets a counsell tax bill is because they've mixed this stuff up. Now ofcourse, if a council sends a bill to somebody who is dead for their counties tax that's funny. If the Americans send a hellfile missile into a house that's not funny. But it's the same kind of estate you make when you start to line on this stuff. ffordd yw'r dyniffog, mae'r dyniffog. Mae'r dyniffog, yna gweithiwch mfawr, yм ychydig i'w ddim yn cwestiynau dyniffog ar dygwith. Chydydig i ddwydo, ddyniffog yn cwestiynau'r dyniffog. Rydyn ni'n gwneud allan ni'n gyllid i'w dynifog. Rydyn ni'n gwneud. Mae'r Dyniffog yn ddelch ar y llythg stackedurau cysyllt yn Gymru. Rydyn ni'n ddysg Lordodd y Dyniffog, i'w dnyt i ddylen i'r dyniffog erbyn. Felly mae'n meddwl am y ffordd roeddwn dod o bobl, a dweud yn dillunod y ddweud. A dyno'n bywch. Dyswn yn yr ystod o ddysgu o ddysgu diolet. Fe wnaeth ocho i arferfyniaid y dystod o bleiswyr cymdeithasol. Mae'n amser ar y glfurddiad. Dysgwyr yn ddet wizardio'r ddysgu. Dwi'n meddwl fel o separatedoedd o ddysgu o ddysgu. ..y'r ddatblygu yn cyd-fazor iawn. Y ddatblygu'r ddod, ddim yn dweud ydych chi'n mollu yn yn ychydig... ..a ddim yn ddod yn ddod yn oed. Y ddatblygu'r ddod yn cyd-fazor iawn. Y ddatblygu'r ddod yn cyd-fazor iawn. Ddod yn cyd-fazor iawn o ddod yn y dyfodd. Ydw'n dysgu'r ddod yn eich cyfnod. Y ddatblygu'r ddod yn y dechreu. ac any processor projection is inherently flawed, it's inherently error-prone. When you have people doing that, they can be aware of their errors. Because AI is not intelligent, it just matches data. It can match streams which aren't entirely true. This is the problem with AI, it fuses data and it isn't necessarily always correct. Why do I know this? Because I do this. This is the process I do. This is the one I produced in 2015 called the Fraggigram. This is what data fusion does. I do data fusion. If you go to my website, you can get this diagram. This is anybody in the fraggian industry in Britain. This is all the politicians and these lines are connecting them either for financial, for PR, for academic, whatever. There's two and a half thousand pieces of information in that diagram. It was out of date about a week after I did it. The idea of AI is, and data fusion, it will be updated instantaneously. Every time a new piece of data comes in, it will be updated. That's how they believe they can foretell the world. Literally, every one of these bubbles, you click on it and you go to another piece of data which tells you what that bubble is. That's how it works. We won't see this tomorrow because this is from the top of the hill opposite Crown. It is the best place to see the site, but we couldn't get the bus up there. Now, we have the communications centre, which is Tower, and this is the intelligence centre. In there, linked to the computers, that's where they run the operations at Crown. With the Joint Intelligence Analysis Complex, what they will be doing is running large amounts of databases. Why? As I said, they're abstracting data. It's huge quantities of data. They can't do that in real time. What they have to do is collect huge amounts of data beforehand for months, if not years. They feed all that data into their big machine learning or AI system, and it learns patterns and it produces those analyses, those lovely pictures, which allow them to view the world. That means that they have to surveil everybody all of the time using as much of the data as they can get. They have to do it, and they can't pick and choose data. They can't decide if we're going to be in Djibouti next week, so we will look at Djibouti now. They have to look at Djibouti all the time, collect data all the time, just in case they need to be in Djibouti. That's the problem with this. They have to collect lots of data, and if you're collecting lots of data all of the time, you need communications links. We know where crap links are because ordinary military tenders are publicly advertised. We know where crap links to Milford Hall, Shelton, Heidwickham, Welford Caution. This is the European North African grid. This is from some documents that came out about 10 or 12 years ago. It's about Siphanet and Niphanet. It's the Defence Communication System. There are these cables that go across the Atlantic. These cables are everywhere. There are a million miles of fiber-optic cable going round the planet right now. This is a book that was published in 1980 called Technology and Political Control. This book describes technology developed to wage campaign, which is neither conventional war nor policing. That is what we have now with drones and with data being used for students to strike and to track people down and follow them or at least have influence. The thing about cyberspace and particularly data-matching and all that data fusion is, it tells you whatever you want to hear. It's something that one of these senate investigation subcommittees looked at 10 years of experience of data fusion and they produced a report in 2012. Their report was quite clear. It has produced no intelligence of any value. It's skewed. We think the people running these centres are telling us lies because we can't validate anything they say. Oh, but we should keep doing it because it answers their question. It gives them some idea that they have agency in the world. There's some lovely reports coming out now about... We're at the birth of AI now. We're at the stage where AI actually means something for the first time since Azimov discussed it in the 1950s. The latest one was IBM's debater. Had a debate with somebody and was judged to have... Well, not one, but one, one and last one, it drew. IBM debater, why does it need to debate? Debate is a way of trying to critically assess evidence and as quickly as possible give a response. Why do they need a machine to do that is because they need to automate the process of data processing. But of course, as Microsoft found out last year, when you plug an AI into Twitter, within a day, it becomes a racist and sojournist. That's because if you feed it a certain type of data, you get the answer that data indicates. We're not dealing with intelligence, we're dealing with patterns. But what 10, 15 years of data fusion in America tells us was some really good works for the ACLU. You get a system that digests large quantities of data as the Senate subcommittee said much of it is erroneous data. But it targets anti-war activists. It targets environmental activists. It targets all those people who the people running the system really don't like. There's this lovely paper that came out a few years ago in a journal called Surveillance in Society, which is not a page turn, but the prominent role of domestic security services over the past generation has been to completely remove the distinction between policing and information collection. Over the past decade, this tendency has become unmistakable as a result of a frenzied privatisation of state security under the guise of home and security. In the bargain, a new agency of political surveillance has arisen the fusion centre. This phenomenon is a medium of low privatisation and assault on ever-sharing civil liberties in an ever more militarised, ever more insecure society. That is the conclusion of that 1980 book, The Technology of Political Control. And it arrived at that conclusion in 1980, not from a detailed analysis of the technology, but from the people who are commissioning it. Because if you build a system, you tend to build your own prejudices into it. Crown is just one of many sites. All sites have been influenced by this data-centric operation idea. They're all becoming more intrusive, they're all becoming more invasive for civil liberties, of privacy and a freedom of expression by the menace that they create in society, the idea that everyone is being watched. The big one is that because these data centres don't work within any more place, they're a network, they can actually route data by a pre-selected set of ideas to avoid certain legal jurisdictions. So they're able to infect me to operate outside the law. And because of this parallel development to the corporate world, there is no distinction between civil and military law. There is no dual use, it's all one machine. And for that reason, as peace activists, you've got to start talking to privacy activists. You've got to start talking to people who worry about freedom of expression. That is how we build the peace and conflict message. Not by trying to find interests amongst people who are interested in peace and conflict, but people who are interested in privacy, in expression, in surveillance, in civil liberties, in democracy. Because that is what this system is there to influence, is there to create. And so what create this to me now, it's an exemplar of the problem of an unaccountable, opaque system of government. And if we would deal with it as that, not as just weapons of war, but actually as a system that doesn't work anymore, as we heard somebody say earlier, it's a system that doesn't work, not democracy, it's the politics that arises from it. Then we can make the peace message via all those other people.