Re Oxford Report: Computational Propaganda Worldwide

["Dave Farber" <dave () farber net>]

---------- Forwarded message ---------
From: Douglas Guilbeault <douglasguilbeault () gmail com>
Date: Tue, Jun 27, 2017 at 10:11 PM
Subject: Re: [IP] Oxford Report: Computational Propaganda Worldwide
To: L Jean Camp <ljeanc () gmail com>
Cc: Dave Farber <dave () farber net>, Sam Woolley <woolleysam () gmail com>, ip <
ip () listbox com>, Douglas Guilbeault <dguilbeault () asc upenn edu>


Thanks for sending. Yes, I wonder if you could measure silencing by
examining anomalies in the burstiness of messaging toward a particular
user. E.g. if a network of bots starting blasting messages at a user around
the same time. Silencing could be detected if a user significantly reduced
their frequency of tweets as a result of bursty attacks from bots. Perhaps
some regression discontinuity analysis would be effective here as a method
for measuring quasi-causality. Perhaps the burstiness wouldn't be essential
- i.e. perhaps you could find that messages in general from bots reduced
tweeting frequency in a targeted user, but it would be tricky to tease out
confounding variables for why a user may fluctuate in tweet frequency (e.g.
just normal fluctuations in normal usage). I am meeting with a friend from
the ADL and Internet & Society to discuss hate speech and how to measure it
online. This may lead to the kind of data that would be ideal for testing
the methods we’re discussing.

On Thu, Jun 22, 2017 at 8:26 PM, L Jean Camp <ljeanc () gmail com> wrote:

> I would be happy to discuss this with you.
>
> I think you could probably an interval techniques, which we used for
> routing,  to look for silencing. It is worth trying, imho.  I realize the
> number are an order of magnitude smaller; however, we did a study where we
> recreated the RIB as a graph and then reconstructed it at steps to
> determine if there were large scale incidents emerging. This reconstruction
> at a time step was the method I was considering when I was talking about
> looking at who disappeared: look at what entities are tweeting for HRC,
> which are identified as bots, and who was not a bot but dropped out. Our
> current submission is attached. But it is about routing, not Twitter. OTOH
> if you used "dropped out" as a measure for bots, you would get the real
> people who were targeted by bots and supported HRC, then equivocate them to
> actual hate bots.
>
> Anyway the method is discussed in the attached.
>
> Prof. L. Jean Camp
> http://www.ljean.com
>
> Make a Difference
> http://www.ieeeusa.org/policy/govfel/congfel.asp
>
> On Thu, Jun 22, 2017 at 6:31 PM, Douglas Guilbeault <
> douglasguilbeault () gmail com> wrote:
>
> > Dear Professor Camp,
> >
> >
> > This is Doug writing on behalf of the computational propaganda team. We
> > really appreciate your note. We agree it is vital that social media is not
> > used to silence women or to promote any kind of hate speech. We also think
> > it is crucial to investigate biases of all kinds in the study and design of
> > algorithms, and we think that the absence of studies on these topics is
> > indicative of a broader issue that needs to be addressed in the field. We
> > are interested in learning about how to explore these issues further.
> >
> >
> > I’m currently working on a chapter for the second edition of Zizi
> > Papacharissi’s, *The Networked Self*. In this chapter, I'd like to
> > discuss some of the issues surrounding bots and discrimination. If you have
> > time this summer to discuss one on one, I would appreciate the opportunity.
> >
> >
> > Sincerely,
> >
> > Doug
> >
> > On Wed, Jun 21, 2017 at 3:41 PM, L Jean Camp <ljeanc () gmail com> wrote:
> >
> > > Silencing disagreement is a critical third role of bots, and they were
> > > effective. This report  adds disappearing the impact of bots and hate
> > > campaigns to the disappearing of voices online.
> > >
> > > I was a member of multiple secret groups on Facebook that existed for
> > > women to express support without facing abuse, threats, and retaliation.
> > > Massive targeted threats of violence and death are silencing. This report
> > > misses a critical experience of Clinton supporters: threats and the
> > > associate silencing. The invisible Clinton supporter was invisible for a
> > > reason. And that reason was silencing.
> > >
> > > I have copied the authors because I do not know if they think silencing
> > > harassment of women is such an online norm it is not worth exploring
> > > further; if they do not think matters; or they do not believe it exists and
> > > so did not look for it.
> > >
> > > You can measure the absence of these voices with the tools that are
> > > being used but YOU HAVE TO LOOK FOR THEM. To my knowledge no one has looked
> > > for the creation and then silencing of pro-Clinton voices as voices spoke
> > > up and then were silenced.
> > >
> > > This matters because women were being silenced with mass attacks and
> > > continue to be silenced.  Not just during the election and by GamerGate,
> > > even earlier.  Any pro-women Usenet group was slammed with jerks explaining
> > > the world to and dumping abuse on women. That is why Systers was created in
> > > the nineties. So it is not as if it is some new thing no one would know to
> > > look for in new media.
> > >
> > > The tools do not generate the hypothesis. You have to look at the data a
> > > different way for the question. Choosing not to look at the participants in
> > > the discussion over time instruments an opinion that this silencing is not
> > > important. Stating that it was not a role of bots makes that explicit.
> > >
> > > I was recently told by a male colleague that he had never had a death
> > > threat online. I thought my teeth would fall out. My first online threats
> > > of murder and sexual assault  were at CMU.  My most recent were when
> > > Hannity mentioned me on his show, but those were mostly general
> > > eliminationist threats not more personal death threats, even if they came
> > > in my personal email.  Interestingly I have not gotten any death threats on
> > > twitter, possibly because I am lower profile and also block at first idiot
> > > contact.
> > >
> > > Why is this invisible to so many people? I do not know. But I know its
> > > invisibility matters. And this report reifies it.
> > >
> > > As an aside, I thought the turn of phrase "democratizing propaganda" was
> > > a bad choice, given the demographics of exclusion.  I guess it is the
> > > post-Voting Rights Act or pre-1921 kind of democratizing.
> > >
> > > Please do not let this disappear the act of silencing.
> > >
> > >
> > >
> > > Prof. L. Jean Camp
> > > http://www.ljean.com
> > >
> > > Make a Difference
> > > http://www.ieeeusa.org/policy/govfel/congfel.asp
> > >
> > > On Wed, Jun 21, 2017 at 8:03 AM, Dave Farber <dave () farber net> wrote:
> > >
> > > > ---------- Forwarded message ---------
> > > > From: Richard Forno <rforno () infowarrior org>
> > > > Date: Wed, Jun 21, 2017 at 7:02 AM
> > > > Subject: Oxford Report: Computational Propaganda Worldwide
> > > > To: Infowarrior List <infowarrior () attrition org>
> > > > Cc: Dave Farber <dave () farber net>
> > > >
> > > >
> > > > Computational Propaganda Worldwide: Executive Summary
> > > >
> > > >
> > > > http://comprop.oii.ox.ac.uk/2017/06/19/computational-propaganda-worldwide-executive-summary/
> > > >
> > > > We’re very excited to announce the launch of our case study series on
> > > > computational propaganda in 9 different countries.
> > > >
> > > > Find the executive summary, written by Sam Woolley and Phil Howard,
> > > > here.
> > > >
> > > > The Computational Propaganda Research Project at the Oxford Internet
> > > > Institute, University of Oxford, has researched the use of social media for
> > > > public opinion manipulation. The team involved 12 researchers across nine
> > > > countries who, altogether, interviewed 65 experts, analyzed tens of
> > > > millions posts on seven different social media platforms during scores of
> > > > elections, political crises, and national security incidents. Each case
> > > > study analyzes qualitative, quantitative, and computational evidence
> > > > collected between 2015 and 2017 from Brazil, Canada, China, Germany,
> > > > Poland, Taiwan, Russia, Ukraine, and the United States.
> > > >
> > > > The reports can be found at the following links:
> > > >
> > > >         • United States
> > > >         • China
> > > >         • Russia
> > > >         • Poland
> > > >         • Brazil
> > > >         • Canada
> > > >         • Germany
> > > >         • Ukraine
> > > >         • Taiwan
> > > >
> > > >
> > > > http://comprop.oii.ox.ac.uk/2017/06/19/computational-propaganda-worldwide-executive-summary/
> > > > Archives <https://www.listbox.com/member/archive/247/=now>
> > > > <https://www.listbox.com/member/archive/rss/247/28942858-f70fe378> |
> > > > Modify
> > > > <https://www.listbox.com/member/?&;>
> > > > Your Subscription | Unsubscribe Now
> > > > <https://www.listbox.com/unsubscribe/?&&post_id=20170621080347:AC506242-5679-11E7-A8C0-A901F2A8EEDA>
> > > > <http://www.listbox.com>
> >
> >
> > --
> > @DzGuilbeault
> > douglasguilbeault.com
> > The Annenberg School for Communication
> > The University Pennsylvania


-- 
@DzGuilbeault
douglasguilbeault.com
The Annenberg School for Communication
The University Pennsylvania



-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/18849915-ae8fa580
Modify Your Subscription: https://www.listbox.com/member/?member_id=18849915&id_secret=18849915-aa268125
Unsubscribe Now: 
https://www.listbox.com/unsubscribe/?member_id=18849915&id_secret=18849915-32545cb4&post_id=20170628051829:BCAE8B00-5BE2-11E7-877D-EAEF8FB7B4F5
Powered by Listbox: http://www.listbox.com