Figure 24 - uploaded by Gianluca Stringhini
Content may be subject to copyright.
Meme used to penalize the public image of Hillary Clinton (who is represented as Medusa, a monster) and enhance that of Donald Trump (presented as Perseus, the hero who beheaded Medusa).
Source publication
Internet memes are increasingly used to sway and manipulate public opinion. This prompts the need to study their propagation, evolution, and influence across the Web. In this paper, we detect and measure the propagation of memes across multiple Web communities, using a processing pipeline based on perceptual hashing and clustering techniques, and a...
Citations
... Taken together, these perspectives underline the important developments within social media environments for purposeful memetic messaging to promote and market political candidates and ideologies. In what follows, we offer a review of previous research on memes used as tools of political communication (Shifman, 2014;Mina, 2019;Wiggins, 2019) writ large but also the historical evolution of memes in U.S. political campaigns (Kreiss & McGregor, 2018;Ross & Rivers, 2017). In addition, we provide critical examples as to how individuals and groups have used memes as part of digital grassroots campaigns (Tran, 2022) and also to view memes as a form of political branding and marketing (Hunting, 2019). ...
... Yet others make a sharper distinction between memes as imitators and memes as part of a larger argumentative discourse. Galipeau (2023, p. 438) writes "memes have the power to constitute political discourse and to be a mirror of the culture from which they stem (Wiggins, 2019)". As its core, politics "consists of responding to conflict with dialogue" (Blattberg, 2001, p. 193). ...
... Tran (2022) found that users were weaponizing political internet memes to influence public opinion in the 2012, 2016 and also the 2020 U.S. presidential election. Tran's (2022) findings support previous arguments made by others (e. g., Foster, 2014;Tran, 2021;Zannettou et al., 2018). ...
This article explores the concept of memetic election cycles, examining how memes have become a central tool in U.S. political campaigns since at least the 2008 Obama campaign. Through a review of key literature on political communication, participatory culture, and digital marketing, the study analyzes the evolution of memes as viral political content, but also as a viable means to mobilize the masses behind increasingly polarized political parties and campaigns. It highlights how campaigns such as those of Obama, Trump, Biden, and Harris leveraged memes not only for voter engagement but also as branding tools that shaped public perception. Using examples from recent election cycles, including the 2016 and 2020 U.S. presidential elections, the article discusses the role of memes as grassroots digital marketing and viral political advertising. Additionally, the research explores the potential influence of memes on voting behavior and the risks of disinformation. The findings suggest that memes serve as a hybrid form of digital folklore and marketing, influencing both electoral discourse and voter behavior.
... The textual content provides essential information, offering context to support the visuals and enhance the credibility of multimodal fake news [2]. It is possible to create manipulated textual information to resemble legitimate news reports that imitate the style and structure of reputable sources [3]. News consumers are often misled by convincing information that combines manipulated text with false images, audio, or video, increasing the impact of fake news. ...
Nowadays, individuals rely predominantly on online social media platforms, news feeds, websites, and news aggregator applications to acquire recent news stories. This trend has resulted in an increase in the number of available social media platforms, online news feeds, and news aggregator applications. These news platforms have been accused of spreading fake news to gain more attention and recognition. Earlier, this misinformation or fake news used to be propagated only in the text form. However, with the advent of technology, now it is spread in multimodal forms, such as images with text, videos, and audio with textual content. Currently, the automatic fake news detection models are focused on high resource languages and superficial output. Social media users need clarity and reasoning when it comes to identifying fake news, rather than just a superficial classification of news as fake. Providing context, reasoning, and explanations can help users understand why certain news is misleading or false. Hence, a multimodal system has to be developed to identify and justify fake news. In this proposed work, we have developed a multimodal fake news system for the Low Resource Language Tamil with reasoning-based explainability. The dataset for this proposed work is retrieved from fact-check websites and official news websites. We have experimented with different combinations of models for visual and text modalities. Further, we integrated LLM-based image descriptions into our model with the text and visual features, resulting in an F1 score of 0.8736. We used the Siamese model to determine the similarity of the news and its image descriptions. Additionally, we conducted error analysis and used explainable artificial intelligence to explore the reasoning behind our model’s predictions. We also present the textual reasoning for the model’s predictions and match them with images.
... For the tourism industry, this could mean a more personalized and seamless user experience. As internet culture has evolved, a form of content known as "memes" has emerged, characterized by humor and cultural references; this viral content is disseminated online rapidly and is continuously evolving and being integrated into popular online culture [7]. Memes are primarily characterized as being "entertaining and memorable" and their format often differs significantly from conventional language used in daily conversation. ...
A novel framework called TravelRAG is introduced in this paper, which is built upon a large language model (LLM) and integrates Retrieval-Augmented Generation (RAG) with knowledge graphs to create a retrieval system framework designed for the tourism domain. This framework seeks to address the challenges LLMs face in providing precise and contextually appropriate responses to domain-specific queries in the tourism field. TravelRAG extracts information related to tourist attractions from User-Generated Content (UGC) on social media platforms and organizes it into a multi-layer knowledge graph. The travel knowledge graph serves as the core retrieval source for the LLM, enhancing the accuracy of information retrieval and significantly reducing the generation of erroneous or fabricated responses, often termed as “hallucinations”. As a result, the accuracy of the LLM’s output is enhanced. Comparative analyses with traditional RAG pipelines indicate that TravelRAG significantly boosts both the retrieval efficiency and accuracy, while also greatly reducing the computational cost of model fine-tuning. The experimental results show that TravelRAG not only outperforms traditional methods in terms of retrieval accuracy but also better meets user needs for content generation.
... The word meme was first introduced as an idea, behavior, or style that disseminates from person to person within a culture [1]. Memes are usually images combined with text descriptions and are used to express ideas such as humor, embarrassment, hate, and even more emotions nowadays. ...
... Since much humor occurs when people's race and gender are involved [1,29], we believe that not only objects in the image should be identified, but also human faces should be focused on. Thus, we compute both face and foreground object proposals and attributes, which help capture subtle humorous contents and appropriate background context of the input meme. ...
With the increasing boom of social media, users can tweet about different events and topics to convey their feelings and emotions. Among these, memes have been gaining popularity over the years. However, it’s insufficient to detect whether a meme is humorous or not with current multimodal models. Based on that, we have done the following work in this paper. For the insufficiency of public datasets on multimodal humor detection on memes, we construct a multi-lingual humor detection dataset called HUMEMES which contains over 5000 thousand memes. Secondly, we propose a multimodal fusion model called CEFM using CLIP encoder for better text and image representation. We use proposal and attribute information to enhance the representation of both modalities. Our model systematically analyzes the local and the global perspective of the input meme and relates it to the background context. This method can better integrate multimodal information and achieves results that exceed the baseline methods. The full codes and dataset are available at https://github.com/gilgamesh-nlp/CEFM.
... The Internet is a breeding ground for hate speech 1,2 . Hateful content and its followers thrive in networks comprising an interconnected web of communities across multiple social media platforms [3][4][5][6] . The communities on each platform (e.g., Telegram Channel, Gab Group, YouTube Channel) show an ideology rooted in hatred and discrimination, and can have anywhere from a few to a few million members. ...
Local or national politics can be a catalyst for potentially dangerous hate speech. But with a third of the world’s population eligible to vote in 2024 elections, we need an understanding of how individual-level hate multiplies up to the collective global scale. We show, based on the most recent U.S. presidential election, that offline events are associated with rapid adaptations of the global online hate universe that strengthens both its network-of-networks structure and the types of hate content that it collectively produces. Approximately 50 million accounts in hate communities are drawn closer to each other and to a broad mainstream of billions. The election triggered new hate content at scale around immigration, ethnicity, and antisemitism that aligns with conspiracy theories about Jewish-led replacement. Telegram acts as a key hardening agent; yet, it is overlooked by U.S. Congressional hearings and new E.U. legislation. Because the hate universe has remained robust since 2020, anti-hate messaging surrounding global events (e.g., upcoming elections or the war in Gaza) should pivot to blending multiple hate types while targeting previously untouched social media structures.
... In the extant scholarly literature, online communities have been examined from various perspectives, including tourism, e.g., refs. [1][2][3][4][5], information systems, e.g., ref. [6], management, e.g., ref. [7], sociology and communication, e.g., ref. [8], psychology, e.g., ref. [9], pedagogy, e.g., ref. [10], and to some extent healthcare, e.g., ref. [11]. ...
The proliferation of social media has transformed how people engage in communication and community building, with platforms like Facebook becoming central to connecting individuals with shared interests. Despite the extensive formation of tourism-oriented online communities on these platforms, there is a notable lack of comprehensive studies examining their structural and managerial dynamics. This study addresses this gap by systematically analyzing fifty international tourism-focused Facebook communities to develop a novel typology based on the nature and type of information shared. The research identifies significant variations in community sizes, engagement levels, and management structures, highlighting that only 6% of these communities qualify as large, with membership exceeding one million. Contrary to common assumptions, a direct link between community size and engagement was not found, with qualitative factors like community purpose and content type being more influential. A notable correlation was observed between the number of administrators and moderators and the member count, emphasizing the importance of effective community governance. The study’s findings contribute to a deeper theoretical understanding of online community dynamics and offer practical implications for tourism marketers and community managers aiming to optimize engagement strategies on social media platforms. The research sets a foundation for future exploration of the interplay between virtual community management and tourism-related discourse.
... Notes 1. https://comic.fosslien.com/post/123653577551, cf., Zannettou, et al., 2018. 2. Myles, et al., 2020 3. The expression "remember the human" appears frequently in rule text, and is also the first point raised in the reddiquette text. ...
The social platform Reddit hosts a set of communities that denounce offensive behavior, invoking scrutiny and shame on (categories of) individuals. Despite varying in their targets, they all promote actionable content to an audience who can view, share and comment on it. These groups allow a global public to air grievances, enabling both accountability and abuse. Following high profile scandals, Reddit routinely sanctions and purges problematic ‘subreddits’. As a matter of self-preservation, subreddits that watch over the public also maintain heightened scrutiny of their own members. Group rules and other prescriptive texts are a means to instill this scrutiny among a broader audience. In analyzing rules and other content management practices in 68 shaming based subreddits, this paper considers how these groups temper platform-based denunciation.
... Optical character recognition (OCR) refers to the task of extracting text within images into machine-encoded text. A model's OCR proficiency is directly related to its ability to access and interpret online information such as infographics, memes, and screenshots of textual conversations, which are prevalent forms of communication and information dissemination online (Zannettou et al., 2018). We use the Hateful Memes and Memotion datasets to evaluate OCR capabilities. ...
... We first preprocess images to remove any solid color framing elements to isolate the base image, then follow Zannettou et al. (2018) and Morina and Bernstein (2022) in extracting templatized memes by running a perceptual hashing algorithm. ...
... In the social computing space, another line of research focuses on understanding how memes originate (Morina and Bernstein, 2022) and spread across platforms (Zannettou et al., 2018). These treat meme templates as discrete tokens. ...
... The matter is further aggravated by visual information, which provides yet another widespread and consequential source of fake news. For instance, fake news stories that include images spread further than those containing only text (Zannettou et al., 2018). ...
