In the old days (1984), we used to call this "thought crime". And what is a reputable university like Cornell doing with such stuff? Might put out of a job some US national security folks I met recently whose task it is to listen to TV, read papers from the -stans (Kazakhstan, etc.) Barry Wellman _____________________________________________________________________ wellman at chass.utoronto.ca http://www.chass.utoronto.ca/~wellman for fun: http://chass.utoronto.ca/oldnew/cybertimes.php _____________________________________________________________________ "Software Being Developed to Monitor Opinions of U.S." New York Times (10/04/06) P. A24; Lipton, Eric The Department of Homeland Security is funding the development of "sentiment analysis" software by a consortium of major universities that uses natural language processing technology to scan foreign publications for negative views on America and its government. The goal of the three-year, $2.4 million grant is to help DHS locate possible dangers to the U.S. The software would provide Homeland Security personnel with instant access to an entire article that contains subversive statements. While efforts have always been made to stay abreast of global opinions of our country, this new technology will make the process far more efficient. Cornell University, the University of Pittsburgh, and the University of Utah are working on the research, which is led by Joe Kielman, who says it could take several years to get the system in place. He says, "We want to understand the rhetoric that is being published and how intense it is, such as the difference between dislike and excoriate." Kielman noted that they are not monitoring U.S.-based news sources. Currently, the system is being fed hundreds of articles published between 2001 and 2002 from a variety of publications and tested on its ability to discern between similar statements. The task of classifying and ranking opinions expressed about America without error has proven quite challenging, says Cornell computer science professor Claire T. Cardie and University of Pittsburgh computer science professor Janyce M. Wiebe. Electronic Privacy Information Center executive director Marc Rotenberg calls the research "really chilling," and compares it to the Defense Department's aborted Total Information Awareness project. He says the research "seems far afield from the mission of homeland security." http://www.nytimes.com/2006/10/04/us/04monitor.html
Talk about the potential for an unnerving amount of false positives (particularly semantical and contextual nuances flagged as suspicious) and the many potential -- dare I say mostly adverse? -- outcomes resulting from these technologies and their analytical limitations. But looking to develop this kind of stuff hardly is surprising, and ties into the whole "one percent doctrine" of threat and vulnerability analysis that's been fostered, if not encouraged, here in DC since 9/11. [dons grumpy political hat] In a country where accusation and assertion has replaced fact in many key decision areas, this is most unsettling -- particularly if critical decisions (or decision-support materials) fail to recognize the limitations of such technologies and/or forgoes any competent human analysis prior to being acted upon. Scary stuff, indeed. [removes grumpy political hat] So, to change the subject completely and avoid accusations of fostering noisy political debate here, I hope y'all had a great time in Brissy last week! -rick On 10/4/06 7:07 PM, "Barry Wellman" <wellman@chass.utoronto.ca> wrote:
In the old days (1984), we used to call this "thought crime".
And what is a reputable university like Cornell doing with such stuff?
Might put out of a job some US national security folks I met recently whose task it is to listen to TV, read papers from the -stans (Kazakhstan, etc.)
I seem to be missing something here. Of all of the things a government could do that would be objectionable, using tools to aggregate and help analyze open source intelligence surely cannot be that evil. I presume that you don't object to governments reading what others have to say about them--this isn't "mind reading," it's "reading." And while analysis of texts certainly requires interpretation by the researcher, I see no particular reason to believe that making use of computer tools to assist in that analysis would necessitate poorer interpretation. I seem to recall a discussion at some point that spoke in fairly positive terms about nVivo, another tool used in open source intelligence. If you are worried that poorly thought out actions may result from good intelligence, that is another issue. It seems that there is a significant breakdown in the process of communicating intelligence analysis. But I think that comparisons to the Total Information Awareness project are extraordinarily counterproductive. I think making use published, open material is an important line of defense for any nation or police force. It is only a "thought crime" if the writers are persecuted for stating it. Otherwise, it's called "listening." Indeed, I see no reason they should limit their analysis to foreign newspapers. Sure, I would love it if they would open up their analysis for public consumption. But besides the closed nature of the results, is there any reason that this should be different from text analysis systems being used to help people keep up on the web today. For example: Google News: http://news.google.com/news Google Zeitgeist: http://www.google.com/press/zeitgeist.html Technorati: http://www.technorati.com/pop/news/ Blogpulse: http://www.blogpulse.com/ Global Attention Profiles: http://h2odev.law.harvard.edu/ezuckerman We Feel Fine: http://wefeelfine.org/ Not to mention the dozens of products designed to map texts (e.g, http://www.leximancer.com/gallery.html). I'm not convinced that providing tools for a government to better understand public discourse is automatically a bad thing. Alex -- // // This email is // [X] assumed public and may be blogged / forwarded. // [ ] assumed to be private, please ask before redistributing. // // Alexander C. Halavais // Social Architect // http://alex.halavais.net //
Personally, the question seems to me to go to why this is funded by the US government at all, and much more importantly why it's funded by "homeland security." ap ---------------------------------------------------------------------- Andrew J Perrin - andrew_perrin@unc.edu - http://perrin.socsci.unc.edu Assistant Professor of Sociology; Book Review Editor, _Social Forces_ University of North Carolina - CB#3210, Chapel Hill, NC 27599-3210 USA New Book: http://www.press.uchicago.edu/cgi-bin/hfs.cgi/00/178592.ctl On Thu, 5 Oct 2006, Alex Halavais wrote:
I seem to be missing something here. Of all of the things a government could do that would be objectionable, using tools to aggregate and help analyze open source intelligence surely cannot be that evil. I presume that you don't object to governments reading what others have to say about them--this isn't "mind reading," it's "reading." And while analysis of texts certainly requires interpretation by the researcher, I see no particular reason to believe that making use of computer tools to assist in that analysis would necessitate poorer interpretation. I seem to recall a discussion at some point that spoke in fairly positive terms about nVivo, another tool used in open source intelligence.
If you are worried that poorly thought out actions may result from good intelligence, that is another issue. It seems that there is a significant breakdown in the process of communicating intelligence analysis. But I think that comparisons to the Total Information Awareness project are extraordinarily counterproductive. I think making use published, open material is an important line of defense for any nation or police force. It is only a "thought crime" if the writers are persecuted for stating it. Otherwise, it's called "listening." Indeed, I see no reason they should limit their analysis to foreign newspapers.
Sure, I would love it if they would open up their analysis for public consumption. But besides the closed nature of the results, is there any reason that this should be different from text analysis systems being used to help people keep up on the web today. For example:
Google News: http://news.google.com/news Google Zeitgeist: http://www.google.com/press/zeitgeist.html Technorati: http://www.technorati.com/pop/news/ Blogpulse: http://www.blogpulse.com/ Global Attention Profiles: http://h2odev.law.harvard.edu/ezuckerman We Feel Fine: http://wefeelfine.org/
Not to mention the dozens of products designed to map texts (e.g, http://www.leximancer.com/gallery.html).
I'm not convinced that providing tools for a government to better understand public discourse is automatically a bad thing.
Alex
-- // // This email is // [X] assumed public and may be blogged / forwarded. // [ ] assumed to be private, please ask before redistributing. // // Alexander C. Halavais // Social Architect // http://alex.halavais.net // _______________________________________________ The air-l@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/
Et Al, This is not a new phenomena. The NSA, CIA , Defense Intelligence, State Department Intelligence agencies and the bulk of the world intellengence organizations have been examining written documents (books, newspapers, news broadcasts, and scholarly papers) for decades. In fact there is a 60s movie on the subject starring a young Robert Redford called the "Day of the Condor" As a recent graduate and a Marine officer I was an analyst for the CIA decades ago. I also agree with the sentiments of Alex. Sam Alex Halavais <halavais@gmail.com> wrote: I seem to be missing something here. Of all of the things a government could do that would be objectionable, using tools to aggregate and help analyze open source intelligence surely cannot be that evil. I presume that you don't object to governments reading what others have to say about them--this isn't "mind reading," it's "reading." And while analysis of texts certainly requires interpretation by the researcher, I see no particular reason to believe that making use of computer tools to assist in that analysis would necessitate poorer interpretation. I seem to recall a discussion at some point that spoke in fairly positive terms about nVivo, another tool used in open source intelligence. If you are worried that poorly thought out actions may result from good intelligence, that is another issue. It seems that there is a significant breakdown in the process of communicating intelligence analysis. But I think that comparisons to the Total Information Awareness project are extraordinarily counterproductive. I think making use published, open material is an important line of defense for any nation or police force. It is only a "thought crime" if the writers are persecuted for stating it. Otherwise, it's called "listening." Indeed, I see no reason they should limit their analysis to foreign newspapers. Sure, I would love it if they would open up their analysis for public consumption. But besides the closed nature of the results, is there any reason that this should be different from text analysis systems being used to help people keep up on the web today. For example: Google News: http://news.google.com/news Google Zeitgeist: http://www.google.com/press/zeitgeist.html Technorati: http://www.technorati.com/pop/news/ Blogpulse: http://www.blogpulse.com/ Global Attention Profiles: http://h2odev.law.harvard.edu/ezuckerman We Feel Fine: http://wefeelfine.org/ Not to mention the dozens of products designed to map texts (e.g, http://www.leximancer.com/gallery.html). I'm not convinced that providing tools for a government to better understand public discourse is automatically a bad thing. Alex -- // // This email is // [X] assumed public and may be blogged / forwarded. // [ ] assumed to be private, please ask before redistributing. // // Alexander C. Halavais // Social Architect // http://alex.halavais.net // _______________________________________________ The air-l@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org Join the Association of Internet Researchers: http://www.aoir.org/ --------------------------------- Do you Yahoo!? Get on board. You're invited to try the new Yahoo! Mail.
There are two major points I want to make here and one minor personal point about myself to set out my biases first. One I have paranoid schizophrenia. There are two rules of thumb with this illness and delusions, one the person believes they are Jesus, two the person feels the CIA, KGB, Mosad, RCMP or place your own state police name here, agency are spying on them. I am this second type and it is only usually by avoid this topic that I do well. That said I spend the last part of my second BA in legal studies studying political criminal law including learning to be critical of the state security agendas. Now my points I wonder if the intelligence agents are reading open source computer code? This would seem to be a good idea given the back room nature of computer security programming in the Unix Linux secure computer paranoia mind set. Mind you that's my personal opinion of computer security geeks their over doing the security. Why is the inventor of PGP cool anyway? Therefore is open source code actually studied by open source intelligence analysts? Before you think I am just being cute or rigidly concrete in connecting the words open source in two contexts, consider that the effects of software can be harmful and socially significant. We all know Microsoft code(s) affect us and courts can open up this code. Will we keep a close eye on open source code? Many former computer students in the world and there are millions can read computer code. My second point is this... from being a long time opponent of big organizations and say something like the welfare payment system being handled by computers and the significant errors funny, absurd, and fatal that occur. In this case of social support systems I know that good information can save lives. Guardians, social workers, self help and other information brokers can help a welfare recipient with the computer systems. Also there is much written in computer ethics about computer error and safety. Thus if we consider reading a benign activity as I believe many academics may naturally do, is it really that safe to have a computer reading for our armed professionals? Just because the tool works is it really error free? What if these data gathering tools make mistakes? i/e did the swat team go to the wrong house because of some open source GIS that some CIA analyst read with the help of a computer that in error added two digits to the zip code? Closing with one more opinion 1984 was written by an intelligence officer for the British and taught as a view of the Soviet union in Canadian high school English class as the novel Animal Farm. But why would this term 1984 be of aid to the British? I think stuff by Orwell is just that propaganda but for the British way of life. Peter Timusk B.Math(2002) BA (2006) Tel: 001-613-729-8328 Community Informatics Practitioner Email: ptimusk@sympatico.ca Yahoo ID: crystal_computing Skype ID: peter.timusk ++++++++++++++++++++++++++++++++++ Nothing I write is intended to be representative of my employer, or our clients. Nor do I alone speak for my unions. ++++++++++++++++++++++++++++++++++ Feel free to learn more about me at www.crystalcomputing.net Computer ethics studies at www.webpagex.org blogs http://logbook.crystalcomputing.net <- computers http://notebook.webpagex.org <- school work On 5-Oct-06, at 12:33 AM, Alex Halavais wrote:
I seem to be missing something here. Of all of the things a government could do that would be objectionable, using tools to aggregate and help analyze open source intelligence surely cannot be that evil. I presume that you don't object to governments reading what others have to say about them--this isn't "mind reading," it's "reading." And while analysis of texts certainly requires interpretation by the researcher, I see no particular reason to believe that making use of computer tools to assist in that analysis would necessitate poorer interpretation. I seem to recall a discussion at some point that spoke in fairly positive terms about nVivo, another tool used in open source intelligence.
If you are worried that poorly thought out actions may result from good intelligence, that is another issue. It seems that there is a significant breakdown in the process of communicating intelligence analysis. But I think that comparisons to the Total Information Awareness project are extraordinarily counterproductive. I think making use published, open material is an important line of defense for any nation or police force. It is only a "thought crime" if the writers are persecuted for stating it. Otherwise, it's called "listening." Indeed, I see no reason they should limit their analysis to foreign newspapers.
Sure, I would love it if they would open up their analysis for public consumption. But besides the closed nature of the results, is there any reason that this should be different from text analysis systems being used to help people keep up on the web today. For example:
Google News: http://news.google.com/news Google Zeitgeist: http://www.google.com/press/zeitgeist.html Technorati: http://www.technorati.com/pop/news/ Blogpulse: http://www.blogpulse.com/ Global Attention Profiles: http://h2odev.law.harvard.edu/ezuckerman We Feel Fine: http://wefeelfine.org/
Not to mention the dozens of products designed to map texts (e.g, http://www.leximancer.com/gallery.html).
I'm not convinced that providing tools for a government to better understand public discourse is automatically a bad thing.
Alex
-- // // This email is // [X] assumed public and may be blogged / forwarded. // [ ] assumed to be private, please ask before redistributing. // // Alexander C. Halavais // Social Architect // http://alex.halavais.net // _______________________________________________ The air-l@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http:// listserv.aoir.org/listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/
Peter Timusk wrote:
Now my points I wonder if the intelligence agents are reading open source computer code? Sure. The NSA even developed its own version of a "Security-Enhanced Linux" (SELinux), which they of course have to give away under the GPL.
See http://www.wired.com/news/linux/0,1411,53004,00.html Ralf -- ----------------------------------------------------- Dipl. Pol. Ralf Bendrath Universität Bremen Sonderforschungsbereich 597 "Staatlichkeit im Wandel" Linzer Str. 9a, D-28359 Bremen, Germany Tel. +49 (421) 218-8735 Fax +49 (421) 218-8721 official http://www.state.uni-bremen.de personal http://userpage.fu-berlin.de/~bendrath blog http://bendrath.blogspot.com
Ralf Bendrath wrote:
Peter Timusk wrote:
Now my points I wonder if the intelligence agents are reading open source computer code?
Sure. The NSA even developed its own version of a "Security-Enhanced Linux" (SELinux), which they of course have to give away under the GPL.
while the german 'intelligence community' brewed its own flavor called 'SINA' (Secure Inter-Networking Architecture), basically a stripped and hardened linux. interesting trivia for geeks: a micro-kernel (!) implementation was developed: http://os.inf.tu-dresden.de/mikrosina/index.xml.en of course all gpl'ed. sina is the backbone of the vpn system used for communication between german embassies. the developers claim it to be the first system that is certified for "top-secret" communication. see: http://www.bsi.de/fachthem/sina/sysbesch/sysbesch.htm (german!) for the german reades here - I published an article 'Electronic Government und die Free Software Bewegung: Der Hacker als Avantgarde Citoyen'* *investigating the gpl-sovereignty nexus in the 2006 book 'politiken der medien', ed. daniel gethman, diaphanes verlag. I am happy to share (and a translation is underway). christoph
See http://www.wired.com/news/linux/0,1411,53004,00.html
Ralf
-- Christoph Engemann, Dipl. Psych. Graduate School of Social Sciences University of Bremen Postfach 330 440 28334 Bremen/Germany Telephone: ++049 179 1233 933 engemann@gsss.uni-bremen.de www.gsss.uni-bremen.de http://cyberlaw.stanford.edu/blogs/engemann/
I wouldn't speak of "thought crime" but of dangerous naivety (if such a behaviour is not intentionally naïve). Training a content-analysis and data mining system working for foreign newspaper articles with US newspaper articles! This sounds surprising , to put it mildly. You understand the Pravda when you know how to read the New York Times? And you use the same technology for understanding both sources? A general solution then is to say: common sense tells everyone that .. So, you would get a common denominator. One might in fact think that there is some common sense in all cultures, and we can base our judgements about the realities (and dangers) of life on common sense, at least. But the last Atlantic Monthly November edition shows that common sense as understood by an American prosecutor is not always the same as the common sense as it is understood by a religious Pakistani: http://www.theatlantic.com/doc/200610/waldman-islam What happens if you fall into the hands of a government which knows visibly nothing of intercultural differences? - Frank Barry Wellman wrote:
In the old days (1984), we used to call this "thought crime".
And what is a reputable university like Cornell doing with such stuff?
Might put out of a job some US national security folks I met recently whose task it is to listen to TV, read papers from the -stans (Kazakhstan, etc.)
Barry Wellman _____________________________________________________________________
wellman at chass.utoronto.ca http://www.chass.utoronto.ca/~wellman for fun: http://chass.utoronto.ca/oldnew/cybertimes.php _____________________________________________________________________
"Software Being Developed to Monitor Opinions of U.S." New York Times (10/04/06) P. A24; Lipton, Eric
The Department of Homeland Security is funding the development of "sentiment analysis" software by a consortium of major universities that uses natural language processing technology to scan foreign publications for negative views on America and its government. The goal of the three-year, $2.4 million grant is to help DHS locate possible dangers to the U.S. The software would provide Homeland Security personnel with instant access to an entire article that contains subversive statements. While efforts have always been made to stay abreast of global opinions of our country, this new technology will make the process far more efficient. Cornell University, the University of Pittsburgh, and the University of Utah are working on the research, which is led by Joe Kielman, who says it could take several years to get the system in place. He says, "We want to understand the rhetoric that is being published and how intense it is, such as the difference between dislike and excoriate." Kielman noted that they are not monitoring U.S.-based news sources. Currently, the system is being fed hundreds of articles published between 2001 and 2002 from a variety of publications and tested on its ability to discern between similar statements. The task of classifying and ranking opinions expressed about America without error has proven quite challenging, says Cornell computer science professor Claire T. Cardie and University of Pittsburgh computer science professor Janyce M. Wiebe. Electronic Privacy Information Center executive director Marc Rotenberg calls the research "really chilling," and compares it to the Defense Department's aborted Total Information Awareness project. He says the research "seems far afield from the mission of homeland security."
http://www.nytimes.com/2006/10/04/us/04monitor.html
_______________________________________________ The air-l@listserv.aoir.org mailing list is provided by the Association of Internet Researchers http://aoir.org Subscribe, change options or unsubscribe at: http://listserv.aoir.org/listinfo.cgi/air-l-aoir.org
Join the Association of Internet Researchers: http://www.aoir.org/
-- .......................................... Dr. Frank Thomas FTR Internet Research 93110 Rosny-sous-Bois France
participants (9)
-
Alex Halavais -
Andrew J Perrin -
Barry Wellman -
Christoph Engemann -
Frank Thomas -
Peter Timusk -
Ralf Bendrath -
Richard Forno -
Sam Tilden