Algorithmic detection of misinformation and disinformation: Gricean perspectives

Published date12 March 2018
DOIhttps://doi.org/10.1108/JD-05-2017-0075
Date12 March 2018
Pages309-332
AuthorSille Obelitz Søe
Subject MatterLibrary & information science,Records management & preservation,Document management,Classification & cataloguing,Information behaviour & retrieval,Collection building & management,Scholarly communications/publishing,Information & knowledge management,Information management & governance,Information management,Information & communications technology,Internet
Algorithmic detection of
misinformation and disinformation:
Gricean perspectives
Sille Obelitz Søe
Department of Information Studies, University of Copenhagen, Copenhagen, Denmark
Abstract
Purpose With the outset of automatic detection of information, misinformation, and disinformation,
the purpose of this paper is to examine and discuss various conceptions of information, misinformation,
and disinformation within philosophy of information.
Design/methodology/approach The examinations are conducted within a Gricean framework in order
to account for the communicative aspects of information, misinformation, and disinformation as well as the
detection enterprise.
Findings While there often is an exclusive focus on truth and falsity as that which distinguish information
from misinformation and disinformation, this paper finds that the distinguishing features are actually
intention/intentionality and non-misleadingness/misleadingness with non-misleadingness/misleadingness
as the primary feature. Further, the paper rehearses the argument in favor of a true variety of disinformation
and extends this argument to include true misinformation.
Originality/value The findings are novel and pose a challenge to the possibility of automatic detection of
misinformation and disinformation. Especially the notions of true disinformation and true misinformation,
as varieties of disinformation and misinformation, which force the true/false dichotomy for information vs
mis-/disinformation to collapse.
Keywords Algorithms, Communication, Information, Philosophy, Misinformation, Disinformation
Paper type Conceptual paper
1. Introduction
The internet is full of communication. Some of this communication consists of false,
inaccurate, and untrue information. As a reaction to this we have lately witnessed an
increased interest in automatic detection (through algorithms) of misinformation and
disinformation. Examples of such detecting-projects are the PHEME-project (2014),
Kumar and GeethakumarisTwitter algorithm(2014), Karlova and Fishers diffusion
model (Karlova and Fisher, 2013), and the Hoaxy platform (Shao et al., 2016) to name a few.
The interest in detection of misinformation and disinformation follows an ancient
philosophical quest for the truth.The hope is to be able to single out misinformation and
disinformation in order to prevent it from spreading, thereby enabling the spread of proper
information –“the truth”–instead.
It is a new task for the algorithmic moderators of social media and the internet.
The assumption is that false content online should be flagged maybe even removed in
order to secure the best conditions for true content, thereby, helping people make the
Journal of Documentation
Vol. 74 No. 2, 2018
pp. 309-332
Emerald Publishing Limited
0022-0418
DOI 10.1108/JD-05-2017-0075
Received 24 May 2017
Revised 6 October 2017
Accepted 17 October 2017
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/0022-0418.htm
© Sille Obelitz Søe. Published by Emerald Publishing Limited. This article is published under the
Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and
create derivative works of this article (for both commercial and non-commercial purposes), subject to
full attribution to the original publication and authors. The full terms of this licence may be seen at
http://creativecommons.org/licences/by/4.0/legalcode
At the authors PhD defenceProfessor David Bawden asked the author:If I am on a deserted island
and your thesisis all I have with me, what would you like me to gain from it? What shouldI remember?
This paper provides the authors answer. Thus, it is based on the authors dissertation (Søe, 2016).
The author would like to thank Jens-Erik Mai, Erik J. Olsson, Jack Andersen, and Laura Skouvig for
valuable comments and suggestions along the way.
309
Gricean
perspectives
right decisions. That is, in crude terms, to decide what people should believe. Further, the
assumption is that the detection of truth and falsity is sufficient for the detection of
misinformation and disinformation. However, the full story is more complicated than mere
detection of truth and falsity which might actually prove complicated enough.
The mechanism of socialmedia is communication. To post and sharestories. To react and
comment. To write statuses about oneself that friends, family, colleagues, and others can
comment upon and share. To arrange and coordinate public and private events and invite
people to attend. Allthese acts both the verbal and non-verbalareactsofcommunication.
They are carried out at a specifictime, within a specific context, and for a specificpurpose
guided by belief, intention, and meaning. When a story is shared the original purpose of
posting it, might change for another purpose insharing it. The context changes as well and,
most likely, also the belief and intention and maybe even the meaning. Thus, to determine
whether something is misinformation or disinformation requires evaluative judgments of
content, context,purpose, etc. and the question is whethersuch judgments can be automated.
Further, the main question is what algorithms should look for in order to detect
misinformation and disinformation i.e. what misinformation and disinformation actually is
in connection to one another and in connection to information. That is, what are the distinct
and distinguishingfeatures of information,misinformation, and disinformation, conceptually?
The paper is structured as follows. Section 2 sets the scene by dealing with four
different detecting-projects. In Section 3 a Gricean framework of meaning, cooperation,
and communication is laid out. Four different accounts of information, misinformation, and
disinformation, all influenced by Grice, are presented and briefly discussed in Section 4.
Discussions of information, misinformation, and disinformation, their nature, and their
implications for automatic detection are carried out in Section 5 and Section 6 provides
concluding remarks.
2. Automatic detection
The PHEME-project sets out to algorithmically detect and categorize rumors in social
network structures (such as Twitter and Facebook) in near real time. The rumors are
mapped according to four categories, which include misinformation, where something
untrue is spread unwittingly; and disinformation, where its done with malicious intent.
(Sheffield News, 2014). The purpose is to help journalists by developing a platform where
stories and rumors can be fact-checked before a story is posted online or sent to print
(PHEME, 2014). Kumar and Geethakumari (2014) propose an algorithm which can detect
and flag whether a tweet is misinformation or disinformation. In their framework
Misinformation is false or inaccurate information, especially that which is deliberately
intended to deceive [and] Disinformation is false information that is intended to mislead,
especially propaganda issued by a government organization to a rival power or the media.
(Kumar and Geethakumari, 2014, p. 3). The purpose of the algorithm is to improve decision
making for individual users by letting the algorithm tell them whether a given tweet is
information, misinformation, or disinformation, and thereby indirectly tell them whether
they should retweet or not. In Karlova and Fishers (2013) diffusion model misinformation
and disinformation extend the concept of information through their informativeness
and misinforming and disinforming function as types of information behavior. Thus,
Karlova and Fisher (2013) define misinformation as inaccurate information and
disinformation as deceptive information and their goal is to better understand how
information spreads and diffuses in online social networks. Hoaxy (Shao et al., 2016) is a
platform for the collection, detection, and analysis of online misinformation and its related
fact-checking efforts.(Shao et al., 2016, p. 745). Hoaxy solely deals with misinformation
defined as false or inaccurate information(Shao et al., 2016, p. 745) with examples such as
rumors, false news, hoaxes and elaborate conspiracy theories (Shao et al., 2016).
310
JD
74,2

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT