Multimodal Content Research Articles

Exploring Students’ Generative AI-Assisted Writing Processes: Perceptions and Experiences from Native and Nonnative English Speakers

AbstractGenerative artificial intelligence (AI) can create sophisticated textual and multimodal content readily available to students. Writing intensive courses and disciplines that use writing as a major form of assessment are significantly impacted by advancements in generative AI, as the technology has the potential to revolutionize how students write and how they perceive writing as a fundamental literacy skill. However, educators are still at the beginning stage of understanding students’ integration of generative AI in their actual writing process. This study addresses the urgent need to uncover how students engage with ChatGPT throughout different components of their writing processes and their perceptions of the opportunities and challenges of generative AI. Adopting a phenomenological research design, the study explored the writing practices of six students, including both native and nonnative English speakers, in a first-year writing class at a higher education institution in the US. Thematic analysis of students’ written products, self-reflections, and interviews suggests that students utilized ChatGPT for brainstorming and organizing ideas as well as assisting with both global (e.g., argument, structure, coherence) and local issues of writing (e.g., syntax, diction, grammar), while they also had various ethical and practical concerns about the use of ChatGPT. The study brought to front two dilemmas encountered by students in their generative AI-assisted writing: (1) the challenging balance between incorporating AI to enhance writing and maintaining their authentic voice, and (2) the dilemma of weighing the potential loss of learning experiences against the emergence of new learning opportunities accompanying AI integration. These dilemmas highlight the need to rethink learning in an increasingly AI-mediated educational context, emphasizing the importance of fostering students’ critical AI literacy to promote their authorial voice and learning in AI-human collaboration.

Multimodal fake news detection through intra-modality feature aggregation and inter-modality semantic fusion

The prevalence of online misinformation, termed “fake news”, has exponentially escalated in recent years. These deceptive information, often rich with multimodal content, can easily deceive individuals into spreading them via various social media platforms. This has made it a hot research topic to automatically detect multimodal fake news. Existing works made a great progress on inter-modality feature fusion or semantic interaction yet largely ignore the importance of intra-modality entities and feature aggregation. This imbalance causes them to perform erratically on data with different emphases. In the realm of authentic news, the intra-modality contents and the inter-modality relationship should be in mutually supportive relationships. Inspired by this idea, we propose an innovative approach to multimodal fake news detection (IFIS), incorporating both intra-modality feature aggregation and inter-modality semantic fusion. Specifically, the proposed model implements a entity detection module and utilizes attention mechanisms for intra-modality feature aggregation, whereas inter-modality semantic fusion is accomplished via two concurrent Co-attention blocks. The performance of IFIS is extensively tested on two datasets, namely Weibo and Twitter, and has demonstrated superior performance, surpassing various advanced methods by 0.6 The experimental results validate the capability of our proposed approach in offering the most balanced performance for multimodal fake news detection tasks.

Towards determining perceived audience intent for multimodal social media posts using the theory of reasoned action

Increasing use of social media has resulted in many detrimental effects in youth. With very little control over multimodal content consumed on these platforms and the false narratives conveyed by these multimodal social media postings, such platforms often impact the mental well-being of the users. To reduce these negative effects of multimodal social media content, an important step is to understand creators’ intent behind sharing content and to educate their social network of this intent. Towards this goal, we propose Intent-o-meter, a perceived human intent prediction model for multimodal (image and text) social media posts. Intent-o-meter models ideas from psychology and cognitive modeling literature, in addition to using the visual and textual features for an improved perceived intent prediction model. Intent-o-meter leverages Theory of Reasoned Action (TRA) factoring in (i) the creator’s attitude towards sharing a post, and (ii) the social norm or perception towards the multimodal post in determining the creator’s intention. We also introduce Intentgram, a dataset of 55K social media posts scraped from public Instagram profiles. We compare Intent-o-meter with state-of-the-art intent prediction approaches on four perceived intent prediction datasets, Intentonomy, MDID, MET-Meme, and Intentgram. We observe that leveraging TRA in addition to visual and textual features—as opposed to using only the latter–results in improved prediction accuracy by up to 7.5%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$7.5\\%$$\\end{document} in Top-1 accuracy and 8%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$8\\%$$\\end{document} in AUC on Intentgram. In summary, we also develop a web browser application mimicking a popular social media platform and show users social media content overlaid with these intent labels. From our analysis, around 70%\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$70\\%$$\\end{document} users confirmed that tagging posts with intent labels helped them become more aware of the content consumed, and they would be open to experimenting with filtering content based on these labels. However, more extensive user evaluation is required to understand how adding such perceived intent labels mitigate the negative effects of social media.

PowerPoint presentations in L1-medium and English-medium lectures: examining knowledge-building practices in multimodal slide content

Research on the effectivity of PowerPoint presentations as an adjunct to theoretical and practical content during university lectures has garnered significant yet inconclusive findings. Specifically, how multimodal academic content should be organised to aid communication remains unclear. Legitimation Code Theory (LCT) introduces the concept of ‘semantic waves’ as an effective means to understand cumulative knowledge-­building practices by modelling how forms of academic content are unpacked and repacked to facilitate understanding. This study applied ‘semantic waves’ to better understand how academic knowledge is constructed in multimodal PowerPoint presentations and whether differences exist based on discipline, language of instruction and modes used. Seventy-two lectures from subjects taught in the L1 and English-medium instruction were examined across four disciplines at two Spanish universities. Results showed that slide content in the soft sciences developed semantic waves to a greater extent, particularly in L1 Arts & Humanities lectures. Limiting slide content to the combination of text and graphics appeared to constrain scaffolding strategies in the process of meaning-making, while video content was linked to semantic waving. Finally, some pedagogical implications based on these results are presented.

EduCross: Dual adversarial bipartite hypergraph learning for cross-modal retrieval in multimodal educational slides

In the digital education landscape, cross-modal retrieval (CMR) from multimodal educational slides represents a significant challenge, particularly because of the complex nature of academic content, which includes images, diagrams, equations, and tables across various subjects such as mathematics and biology. Current CMR systems are primarily designed for “(natural) image to text” interactions (or vice versa) and inadequately address real-world educational scenarios. This study presents EduCross, a novel framework devised to enhance CMR within multimodal educational slides, which is a domain in which traditional retrieval systems fall short. Recognizing the imperative for a system that is tailored to the educational context, EduCross integrates dual adversarial bipartite hypergraph learning, harnessing the capabilities of generative adversarial networks with figure-text dual channels. This powerful combination facilitates robust bidirectional mapping, allowing for the precise association of figures with their descriptive spoken language segments and ensuring a comprehensive CMR experience. Specifically, we develop framelet-based deep bipartite hypergraph neural networks that effectively manage the high-order relationships between diverse educational content types and various types of slide figures. Our experimental results underscore the superior performance of EduCross, demonstrating its effectiveness through the use of the real Multimodal Lecture Presentations dataset that mirrors authentic educational settings. These outcomes highlight the significant advancements of EduCross over existing methods, marking a leap forward in the accurate retrieval of multimodal educational content.

Expansive assessment of expansive abilities: Teachers’ perspectives and practices with multimodal and translanguaged content assessments

BackgroundExpansive assessment seeks to reimagine assessment in content and language integrated learning (CLIL) through the lens of expansive theories of communication (e.g., multimodality, translanguaging). However, for expansive assessment to be consequential for students in CLIL classrooms, teachers must be open to recognizing student learning expressed in ways that go beyond what has been traditionally privileged (e.g., written English). AimsThis article reports on two complementary studies of U.S. teachers’ perspectives and practices related to expansive assessment. Study 1 addressed multimodal assessment with pre-service teachers (PSTs) in science. Study 2 addressed translanguaged assessment with in-service teachers (ISTs) in language arts. SamplesParticipants in Study 1 were 10 PSTs. Participants in Study 2 were two ISTs from different elementary schools with different instructional arrangements. MethodsInterviews and observations were used to elicit teachers' perspectives on expansive assessment and their practices interpreting students’ performances. ResultsIn both studies, teachers' perspectives closely reflected their practices. For example, PSTs whose perspectives eschewed logocentrism tended to recognize evidence of science learning in students' visual responses. However, there were also findings unique to each study. For example, ISTs’ perspectives and practices related to translanguaged assessment were mediated by the language policies at their schools. ConclusionsThe two studies, which examined different teacher populations, assessment approaches, and content areas, converge in highlighting three overarching themes to guide future research: (a) the importance of clarity regarding constructs assessed, (b) the contextual mediation of teachers’ perspectives and practices, and (c) the need for teacher learning across their careers.

MIMIC: Misogyny Identification in Multimodal Internet Content in Hindi-English Code-Mixed Language

Over the years, social media has emerged as one of the most popular platforms where people express their views and share thoughts about various aspects. The social media content now includes a variety of components such as text, images, videos etc. One type of interest is memes, which often combine text and images. It is relevant to mention here that, social media being an unregulated platform, sometimes also has instances of discriminatory, offensive and hateful content being posted. Such content adversely affects the online well-being of the users. Therefore, it is very important to develop computational models to automatically detect such content so that appropriate corrective action can be taken. Accordingly, there have been research efforts on automatic detection of such content focused mainly on the texts. However, the fusion of multimodal data (as in memes) creates various challenges in developing computational models that can handle such data, more so in the case of low-resource languages. Among such challenges, the lack of suitable datasets for developing computational models for handling memes in low-resource languages is a major problem. This work attempts to bridge the research gap by providing a large-sized curated dataset comprising 5,054 memes in Hindi-English code-mixed language, which are manually annotated by three independent annotators. It comprises two subtasks: (i) Subtask-1 (Binary classification involving tagging a meme as misogynous or non-misogynous), and (ii) Subtask-2 (multi-label classification of memes into different categories). The data quality is evaluated by computing Krippendorff's alpha. Different computational models are then applied on the data in three settings: text-only, image-only, and multimodal models using fusion techniques. The results show that the proposed multimodal method using the fusion technique may be the preferred choice for the identification of misogyny in multimodal Internet content and that the dataset is suitable for advancing research and development in the area.

IMF-MF: Interactive moment localization with adaptive multimodal fusion and self-attention

Precise video moment retrieval is crucial for enabling users to locate specific moments within a large video corpus. This paper presents Interactive Moment Localization with Multimodal Fusion (IMF-MF), a novel interactive moment localization with multimodal fusion model that leverages the power of self-attention to achieve state-of-the-art performance. IMF-MF effectively integrates query context and multimodal features, including visual and audio information, to accurately localize moments of interest. The model operates in two distinct phases: feature fusion and joint representation learning. The first phase dynamically calculates fusion weights for adapting the combination of multimodal video content, ensuring that the most relevant features are prioritized. The second phase employs bi-directional attention to tightly couple video and query features into a unified joint representation for moment localization. This joint representation captures long-range dependencies and complex patterns, enabling the model to effectively distinguish between relevant and irrelevant video segments. The effectiveness of IMF-MF is demonstrated through comprehensive evaluations on three benchmark datasets: TVR for closed-world TV episodes and Charades for open-world user-generated videos, DiDeMo dataset, Open-world, diverse video moment retrieval dataset. The empirical results indicate that the proposed approach surpasses existing state-of-the-art methods in terms of retrieval accuracy, as evaluated by metrics like Recall (R1, R5, R10, and R100) and Intersection-of-Union (IoU). The results consistently demonstrate IMF-MF’s superior performance compared to existing state-of-the-art methods, highlighting the benefits of its innovative interactive moment localization approach and the use of self-attention for feature representation and attention modeling.

NExT-OOD: Overcoming Dual Multiple-Choice VQA Biases.

In recent years, multiple-choice Visual Question Answering (VQA) has become topical and achieved remarkable progress. However, most pioneer multiple-choice VQA models are heavily driven by statistical correlations in datasets, which cannot perform well on multimodal understanding and suffer from poor generalization. In this paper, we identify two kinds of spurious correlations, i.e., a Vision-Answer bias (VA bias) and a Question-Answer bias (QA bias). To systematically and scientifically study these biases, we construct a new video question answering (videoQA) benchmark NExT-OOD in OOD setting and propose a graph-based cross-sample method for bias reduction. Specifically, the NExT-OOD is designed to quantify models' generalizability and measure their reasoning ability comprehensively. It contains three sub-datasets including NExT-OOD-VA, NExT-OOD-QA, and NExT-OOD-VQA, which are designed for the VA bias, QA bias, and VA&QA bias, respectively. We evaluate several existing multiple-choice VQA models on our NExT-OOD, and illustrate that their performance degrades significantly compared with the results obtained on the original multiple-choice VQA dataset. Besides, to mitigate the VA bias and QA bias, we explicitly consider the cross-sample information and design a contrastive graph matching loss in our approach, which provides adequate debiasing guidance from the perspective of whole dataset, and encourages the model to focus on multimodal contents instead of spurious statistical regularities. Extensive experimental results illustrate that our method significantly outperforms other bias reduction strategies, demonstrating the effectiveness and generalizability of the proposed approach.

Dual-view multi-modal contrastive learning for graph-based recommender systems

Personalized recommender systems play a crucial role in various online content-sharing platforms (e.g., TikTok). The learning of representations for multi-modal content is pivotal in current graph-based recommender systems. Existing works aim to enhance recommendation accuracy by leveraging multi-modal features (e.g., image, sound, text) as side information for items. However, this approach falls short in fully discerning users' fine-grained preferences across different modalities. To tackle this limitation, this paper introduces the Dual-view Multi-Modal contrastive learning Recommendation model (DMM-Rec). DMM-Rec employs self-supervised learning to guide the learning of user and item representations within the multi-modal context. Specifically, to capture users' preferences for different modalities, we propose specific-modal contrastive learning. Simultaneously, to capture users' cross-modal preferences, cross-modal contrastive learning is introduced to uncover interdependencies in users' preferences across modalities. The contrastive learning tasks not only adaptively explore potential relations between modalities but also address the data sparsity challenge in recommender systems. Extensive experiments conducted on three datasets and compared against ten baselines demonstrate that DMM-Rec outperforms the strongest baseline by an average of 6.81%. These results underscore the effectiveness of considering multi-modal content in improving recommender systems.

Frequency Spectrum Is More Effective for Multimodal Representation and Fusion: A Multimodal Spectrum Rumor Detector

Multimodal content, such as mixing text with images, presents significant challenges to rumor detection in social media. Existing multimodal rumor detection has focused on mixing tokens among spatial and sequential locations for unimodal representation or fusing clues of rumor veracity across modalities. However, they suffer from less discriminative unimodal representation and are vulnerable to intricate location dependencies in the time-consuming fusion of spatial and sequential tokens. This work makes the first attempt at multimodal rumor detection in the frequency domain, which efficiently transforms spatial features into the frequency spectrum and obtains highly discriminative spectrum features for multimodal representation and fusion. A novel Frequency Spectrum Representation and fUsion network (FSRU) with dual contrastive learning reveals the frequency spectrum is more effective for multimodal representation and fusion, extracting the informative components for rumor detection. FSRU involves three novel mechanisms: utilizing the Fourier transform to convert features in the spatial domain to the frequency domain, the unimodal spectrum compression, and the cross-modal spectrum co-selection module in the frequency domain. Substantial experiments show that FSRU achieves satisfactory multimodal rumor detection performance.

Reinforced Adaptive Knowledge Learning for Multimodal Fake News Detection

Nowadays, detecting multimodal fake news has emerged as a foremost concern since the widespread dissemination of fake news may incur adverse societal impact. Conventional methods generally focus on capturing the linguistic and visual semantics within the multimodal content, which fall short in effectively distinguishing the heightened level of meticulous fabrications. Recently, external knowledge is introduced to provide valuable background facts as complementary to facilitate news detection. Nevertheless, existing knowledge-enhanced endeavors directly incorporate all knowledge contexts through static entity embeddings, resulting in the potential noisy and content-irrelevant knowledge. Moreover, the integration of knowledge entities makes it intractable to model the sophisticated correlations between multimodal semantics and knowledge entities. In light of these limitations, we propose a novel Adaptive Knowledge-Aware Fake News Detection model, dubbed AKA-Fake. For each news, AKA-Fake learns a compact knowledge subgraph under a reinforcement learning paradigm, which consists of a subset of entities and contextual neighbors in the knowledge graph, restoring the most informative knowledge facts. A novel heterogeneous graph learning module is further proposed to capture the reliable cross-modality correlations via topology refinement and modality-attentive pooling. Our proposal is extensively evaluated over three popular datasets, and experimental results demonstrate the superiority of AKA-Fake.

THGFormer: Time-Aware Hypergraph Learning for Multimodal Social Media Popularity Prediction (Student Abstract)

Social media popularity prediction of multimodal user-generated content (UGC) is a crucial task for many real-world applications. However, existing efforts are often limited by missing inter-instance correlations and UGC temporal patterns. To address these issues, we propose a novel time-aware hypergraph Transformer framework, THGFormer. It fully represents inter-instance and intra-instance relations by hypergraphs, captures the temporal dependencies with a time encoder, and enhances UGC's representations via a neighborhood knowledge aggregation. Extensive experiments conducted on two real-world datasets demonstrate that THGFormer outperforms state-of-the-art popularity prediction models across several settings.

Unveiling Implicit Deceptive Patterns in Multi-Modal Fake News via Neuro-Symbolic Reasoning

In the current Internet landscape, the rampant spread of fake news, particularly in the form of multi-modal content, poses a great social threat. While automatic multi-modal fake news detection methods have shown promising results, the lack of explainability remains a significant challenge. Existing approaches provide superficial explainability by displaying learned important components or views from well-trained networks, but they often fail to uncover the implicit deceptive patterns that reveal how fake news is fabricated. To address this limitation, we begin by predefining three typical deceptive patterns, namely image manipulation, cross-modal inconsistency, and image repurposing, which shed light on the mechanisms underlying fake news fabrication. Then, we propose a novel Neuro-Symbolic Latent Model called NSLM, that not only derives accurate judgments on the veracity of news but also uncovers the implicit deceptive patterns as explanations. Specifically, the existence of each deceptive pattern is expressed as a two-valued learnable latent variable, which is acquired through amortized variational inference and weak supervision based on symbolic logic rules. Additionally, we devise pseudo-siamese networks to capture distinct deceptive patterns effectively. Experimental results on two real-world datasets demonstrate that our NSLM achieves the best performance in fake news detection while providing insightful explanations of deceptive patterns.

CAMEL: Capturing Metaphorical Alignment with Context Disentangling for Multimodal Emotion Recognition

Understanding the emotional polarity of multimodal content with metaphorical characteristics, such as memes, poses a significant challenge in Multimodal Emotion Recognition (MER). Previous MER researches have overlooked the phenomenon of metaphorical alignment in multimedia content, which involves non-literal associations between concepts to convey implicit emotional tones. Metaphor-agnostic MER methods may be misinformed by the isolated unimodal emotions, which are distinct from the real emotions blended in multimodal metaphors. Moreover, contextual semantics can further affect the emotions associated with similar metaphors, leading to the challenge of maintaining contextual compatibility. To address the issue of metaphorical alignment in MER, we propose to leverage a conditional generative approach for capturing metaphorical analogies. Our approach formulates schematic prompts and corresponding references based on theoretical foundations, which allows the model to better grasp metaphorical nuances. In order to maintain contextual sensitivity, we incorporate a disentangled contrastive matching mechanism, which undergoes curricular adjustment to regulate its intensity during the learning process. The automatic and human evaluation experiments on two benchmarks prove that, our model provides considerable and stable improvements in recognizing multimodal emotion with metaphor attributes.

An inter-modal attention-based deep learning framework using unified modality for multimodal fake news, hate speech and offensive language detection

Fake news, hate speech and offensive language are related evil triplets currently affecting modern societies. Text modality for the computational detection of these phenomena has been widely used. In recent times, multimodal studies in this direction are attracting a lot of interests because of the potentials offered by other modalities in contributing to the detection of these menaces. However, a major problem in multimodal content understanding is how to effectively model the complementarity of the different modalities due to their diverse characteristics and features. From a multimodal point of view, the three tasks have been studied mainly using image and text modalities. Improving the effectiveness of the diverse multimodal approaches is still an open research topic. In addition to the traditional text and image modalities, we consider image–texts which are rarely used in previous studies but which contain useful information for enhancing the effectiveness of a prediction model. In order to ease multimodal content understanding and enhance prediction, we leverage recent advances in computer vision and deep learning for these tasks. First, we unify the modalities by creating a text representation of the images and image–texts, in addition to the main text. Secondly, we propose a multi-layer deep neural network with inter-modal attention mechanism to model the complementarity among these modalities. We conduct extensive experiments involving three standard datasets covering the three tasks. Experimental results show that detection of fake news, hate speech and offensive language can benefit from this approach. Furthermore, we conduct robust ablation experiments to show the effectiveness of our approach. Our model predominantly outperforms prior works across the datasets.

Interpreting intercepted communication: From talk to evidence

This paper examines the multi-layered translational process of intercepted conversations which appear as evidence in court files in the form of translated wiretap records. The translational challenge here is to transfer the multimodal content of a spoken text in the source language into a written text in the target language. Based on audio data from 17 original intercepted communications and case file documents from a Swiss criminal investigation, the multimodality of the resulting hybrid translational action is explored. In this paper, the process of interlingual decontextualisation and recontextualisation of intercepted communication is examined from the bottom up. The analysis shows that there are different levels of contextualisation that must be considered to achieve a fuller picture of the meaning of a wiretapped conversation.

“Other Customer” Perception as Strategic Insight into Gen Z Consumer–Brand Identification and Purchase Behavior: A Mixed-Methods Approach

This study addresses the knowledge and research gaps regarding the “other customer” concept within the consumer–brand identification paradigm, drawing on branding and consumer behavior theories. We consider the “other customer” as the other customers of a brand from the perspective of an individual (focal customer) using or simply assessing the brand. We take into account both how the focal customer is influenced by other customers and how the focal customer perceives those other customers. While the concept is traditionally associated with the service industries, we expand its relevance to product brands. We use Social Identity Theory, Social Comparison Theory, and Other Customer Perception to identify how Gen Z focal consumers construct the image of other consumers of international brands at national level. A mixed-method approach was developed, bringing together MMCA (multimodal content analysis), perceptual mapping regarding brand image, and other customer perceived attributes, as well as a survey addressing (a) focal customer–other customer–brand identification perceived similarity and (b) perceived influence of sustainability on purchase behavior. MMCA was performed on a corpus of 236 other customer profiles of Nike and Adidas, consisting of a visual component and a textual description. The profiles were developed by a convenience sample of 147 Romanian Gen Z young adults. The results show that: (a) The national specificity and history of the brand can lead to diverging, local other customer profiles that might be accepted or rejected based on the desirability and similarity perceived by the focal customer; (b) The other customer is used as a form of self-evaluation carried out by the focal customer (Gen Z consistently evaluates downward some of the other customers as a way to curate self-esteem and avoid identification with profiles they consider undesirable or unsuitable for their own self-image); and (3) The connection perceivable at international level between brand sustainability as a purchase pre-condition and Gen Z customers is not universal and should be re-evaluated based on local realities. At theoretical level, this study contributes to the advancement of the “other customer” concept; at a practical level, it advocates for the inclusion of an “other customer communication strategy” into the marketing communication mix.

Strengthening inclusive teaching with a MOOC: Adult education and STEM as professional development partners

AbstractThis article explores the design and development of a Massive Open Online Course (MOOC) on inclusive teaching. Released in the Fall of 2023, the MOOC is particularly unique because of its intentional use of ethnically diverse voices and multimodal content to include art, music, spoken word, and storytelling. Modules range from Self and Empathy to Classroom Climate and Networks. MOOC production was graciously funded by a grant from the Howard Hughes Medical Institute. At its core, the article will share reflections on the rationale and process, describe how STEM and Adult Education scholar‐practitioners collaborated to infuse adult education and interdisciplinary concepts in MOOC development and outline meaningful implications for the field.

GenAI against humanity: nefarious applications of generative artificial intelligence and large language models

Generative Artificial Intelligence (GenAI) and Large Language Models (LLMs) are marvels of technology; celebrated for their prowess in natural language processing and multimodal content generation, they promise a transformative future. But as with all powerful tools, they come with their shadows. Picture living in a world where deepfakes are indistinguishable from reality, where synthetic identities orchestrate malicious campaigns, and where targeted misinformation or scams are crafted with unparalleled precision. Welcome to the darker side of GenAI applications. This article is not just a journey through the meanders of potential misuse of GenAI and LLMs, but also a call to recognize the urgency of the challenges ahead. As we navigate the seas of misinformation campaigns, malicious content generation, and the eerie creation of sophisticated malware, we’ll uncover the societal implications that ripple through the GenAI revolution we are witnessing. From AI-powered botnets on social media platforms to the unnerving potential of AI to generate fabricated identities, or alibis made of synthetic realities, the stakes have never been higher. The lines between the virtual and the real worlds are blurring, and the consequences of potential GenAI’s nefarious applications impact us all. This article serves both as a synthesis of rigorous research presented on the risks of GenAI and misuse of LLMs and as a thought-provoking vision of the different types of harmful GenAI applications we might encounter in the near future, and some ways we can prepare for them.