Audio misinformation on WhatsApp: A case study from Lebanon

Abstract

Since 2019, Lebanon has witnessed sequential crises that have routinely spurred media attention. A great deal of misinformation has proliferated during these events, much of it spreading on WhatsApp. One format is particularly understudied: audio instant messages, otherwise known as voice notes. Utilizing a grounded theory approach to examine 35 misleading WhatsApp voice notes collected between October 2019 and October 2020, this study documents how audio misinformation on Lebanese WhatsApp follows a consistent structure through the manipulation of interpersonal relationships, the establishing of source credibility, the imbuing of negative discrete emotions, and the inclusion of calls to action.

Navigating misinformation in voice messages: Identification of user‐centered features for digital interventions

“…While many features can be derived from text as transcribed voice messages, others are based on the audio modality. Research has found that speech rate and sound volume may be used as an indicator when assessing the credibility of audio material (El-Masri et al, 2022)-an insight we build on in our user study.…”

“…While we included a broad variety of features on character, word, and sentence level, this selection is not exhaustive and may be extended based on a systematic literature review, especially on the explicit speech modality that may imbue emotions. Existing research indicates that speech rate and sound volume may compose a relevant attribute of voice messages when assessing its credibility (El-Masri et al, 2022). While these features are not yet included in the prototype implementation, they were part of our user study.…”

“…Specifically, we included whether there was a "claim to be an expert" (e.g.,"I have researched that!") (El-Masri et al, 2022) or a "call for action" (e.g., "Share this voice message!") (El-Masri et al, 2022;Maros et al, 2021).…”

“…(El-Masri et al, 2022) or a "call for action" (e.g., "Share this voice message!") (El-Masri et al, 2022;Maros et al, 2021). Additionally, we considered the existing research on the relevance of creator's profiles and thus included whether a voice message was sent by a "suspicious profile (e.g., name, description of the creator)" which might be detectable via keywords or the excessive usage of emojis.…”

The Networked Trolling of Critical Journalists and News Organizations in Iraq

“…Also, Yaseen and Ali (2022) used discourse analysis to analyze 13 fabricated social media images previously identified by the fact-checking organization Tech4Peace. If we, however, focus on the larger Arab and Middle Eastern context, then we can find more conceptual and empirical studies on misinformation and the reputational attacks against Arab journalists (Al-Rawi 2021a;Bebawi 2022;El-Masri et al 2022;Jones 2021;Mutsvairo and Bebawi 2019) as well as human right activists (Abrahams and Leber 2021;Al-Rawi 2014;Jones 2022). The large majority of the above studies relied on one research method approach, while this paper employs a mixed method approach to identify and analyze disinformation targeting journalists, news organizations, and human rights activists.…”