Digital Age in Semiotics & Communication

A step-by-step semiotic understanding of LLMs and chatbots through interdisciplinary dialogue

Mattia Thibault, Daniele Allois — Tue, 30 Dec 2025 00:00:00 +0000

Artificial Intelligence (AI) is enjoying a period of “summer”. It is an intensely discussed and researched topic which fascinates researchers across fields with questions about both its development and possible impacts. Semiotics too has immediately engaged with the topic, focusing especially on the ways in which different forms of Generative AI can be understood through existing frameworks, and how the processes of textual production it encompasses can be studied and analysed. Gen AI, however, is a complex and fast evolving technology, which is very difficult to study without a sufficient understanding of the technical side. Perceptrons, Artificial Neural Networks, Transformer Networks, Large Language Models are all terms we hear often in discussions about AI. However, they often refer to technologies we barely understand. This paper is born from a collaboration between authors with expertise in semiotics and machine learning. Our objective is to reconstruct the evolution of Large Language Models in particular, and to engage the different stages from a semiotic perspective. This in-depth engagement with the technological side is especially useful to go beyond the strong temptation to anthropomorphize the technology and instead reposition it as a tool that supports human textual creation. Based on our analysis we then propose several key concepts (an AI semiosphere, Generative AI as a mode of sign production and the concept of Intentio Machinae) as ways to map and conceptualise the role of LLMs in semiotic processes.

The Peircean theory of AI: Advancing text generation through Peirce’s triadic model, speculative grammar, and methodeutics

Alon Friedman, Martin Thellefsen — Tue, 30 Dec 2025 00:00:00 +0000

Large language models (LLMs) have significantly advanced natural language processing (NLP) in text generation, translation, and automated question answering. However, despite these advancements, their capacity for interpretative reasoning remains limited. Current AI systems, primarily grounded in formal linguistics and statistical approaches, struggle to capture the relational and contextual dimensions crucial for human-like comprehension. These limitations are particularly evident when interpreting meaning within dynamic social contexts, highlighting the need for theoretical frameworks that extend beyond statistical pattern recognition. This study examines how Charles Sanders Peirce’s nineteenth- century semiotic theory, specifically his triadic model of Sign, Object, and Interpretant, can inform and enhance AI’s interpretative capabilities. Peirce’s systematic approach to meaning making, which predates computational thinking by nearly a century, offers critical insights into the limitations of AI systems grounded primarily in formal logic and statistical operations. These limitations become particularly clear when examining semiotic relationships through the lenses of speculative grammar and methodeutics. Furthermore, we incorporate Claudio Paolucci’s perspective on machinic enunciation and the “myth of meaning” to expand our theoretical framework. Paolucci’s analysis of generative AI as a language-endowed machine, lacking subjectivity yet producing contextually significant enunciates, supports the reinterpretation of AI output in functional and relational terms. This perspective aligns with Peirce’s focus on the triadic process of semiosis, adding a contemporary lens which emphasizes the functional rather than essentialist nature of meaning-making in AI systems. By addressing how Peirce’s triadic model and Paolucci’s framework can bridge the gap between statistical and socially oriented approaches, we contend that Peircean principles can enhance relational understanding in language models and illuminate the theoretical and practical challenges of integrating nineteenth-century semiotic theories into modern computational systems. Our findings indicate that Peirce’s sign theory significantly expands the contextual awareness of AI, highlighting the complexities of replicating interpretative processes. This research demonstrates the continued relevance of classical philosophical frameworks in addressing contemporary technological challenges and contributes to a comprehensive theory of AI.

On the edge of understanding: ChatGpt and the limits of artificial sense-making

Fabio Montesanti — Tue, 30 Dec 2025 00:00:00 +0000

The future challenge of generative AI such as ChatGpt will be to be as similar as possible to a human speaker. Certain notions such as linguistic sentiment and linguistic value in Saussure or decipherer in Jakobson will show how, although the technological evolution of AI is undisputable, the future of AI is more complex than it seems. The human speaker “lives” language – or probably lives within language – while AIs “train themselves” to recognise it mechanically. A fundamental aspect which connotes the human speaker is the adaptability to the communicative situation. During a dialogue, two or more speakers stipulate “a dynamic pact”, mutually adjusting and cooperating to render communication effective and to generate sense. The adaptability to the communicative situation implies the recognition of certain extralinguistic factors (distraction, ambiguity between two different signifiers, the request for confirmation of understanding, etc.). ments which constitute Hymes’ taxonomy, especially the environmental context (spatio-temporal definition of the communicative situation) and the scene (cultural definition of the communicative situation). According to this idea, AIs reproduce a communicative exchange – by generating predetermined messages, through a collective mind, following Lotman – whereas human speakers produce it through their individual minds. From Lotman’s perspective, sense-generating structures rely on a dynamic interplay between symmetry and asymmetry, continually entering into relationship with both synchronic and diachronic structures; in the full article, I will analyze selected dialogues with ChatGPT to explore the potential coexistence of artificial language and natural language, considered by Lotman as a crucial component in the process of sense-making. In these terms, if generative AIs are considered capable of generating sense, then probably the concept of sense – and mind (?) – itself can be semiotically redefined.

The semiotic of AI images

Stephanie Schneider — Tue, 30 Dec 2025 00:00:00 +0000

In her 1942 book Philosophy in a New Key, A Study in the Symbolism of Reason, Rite, and Art, American philosopher Susanne K. Langer distinguishes between two symbolic forms: discursive and presentational. The former is linear, discrete, and successive, while the latter is simultaneous and relational. She describes language as a discursive form and images as presentational. Langer argues that not all meaning can be communicated by discursive symbols; in particular, emotions cannot be expressed in discrete form. Instead, they find their symbolic form in artworks. This aligns with her aesthetic theory, which she elaborates on in her book Feeling and Form (1953), where she states her main proposition that art gives form to our feelings. Her distinction between the representational capabilities of language and images is well-suited for analyzing AI-generated images. Based on Langer, we can see two symbolic forms–the discursive and the presentational–collide in the process of text-to-image generation. Here, the image, as a presentational form, is created from a discursive form, i.e. a description (prompt), which, by Langer’s definition, is incapable of communicating the same meaning as a presentational symbol. This article will explore the Semiotic of AI image generation based on her theory. The limitations of language as a discrete and linear form, and the resulting communicative constraints, are contrasted with images as presentational forms, such as artworks. This is not an evaluation of AI images, but rather an attempt to understand their structure and function within communication processes. The aim is to gain a new perspective on this medium and assess whether these images can help us “make our ideas clear“ (Charles Peirce 1878).

The Janusian face of facial recognition, part 1: Its subface, interface, and surface

Devon Schiller — Tue, 30 Dec 2025 00:00:00 +0000

Facial recognition has a Janusian face. In the instant of its interaction, facial recognition brings many faces together into relation. It is not only visually representational but also computationally re-presentational. To make facial recognition knowable, therefore, one needs first to make its many faces visible. One must expose, so to say, the functional relationalities between rhizomatic facialities within these commercial products and their proprietary computing. Only then can the ways facial recognition technology systems work, and how they are either adversely misused or beneficially used, be substantively challenged or tactically critiqued. Toward this end, from the critical standpoint of a computational semiotics both Peircean and pragmatist, I apply a method of semiotic deblackboxing. I divide my inquiry between two distinct investigations: here in part one, I explicate this theoretical approach; then in part two, I explore its practical application. Across these two parts, I probe how, and in what ways, the artificial intelligence of facial recognition not only comprises a computational, mechanical, and technological system, but also constitutes a semiotic system. The question is this: Does the artificial intelligence system of a facial recognition technology have semiotic agency? That is, is it therefore able to process the action of a sign that has genuine triadic sign relations rather than non-genuine dyadic quasi-sign relations? In other words, is artificial intelligence actually either artificial or intelligent? As I argue here, this Janusian or multiple relation is the intelligent result, not of the semiotic machine, its efficient causation, and allopoiesis, but of the semiotic animal, our final causation, and autopoiesis.

Are we human or are we dancer? – AI creativity in the realm of XR

Momchil Alexiev — Tue, 30 Dec 2025 00:00:00 +0000

As an emerging field of research, the intersection of AI and XR is expanding s in front of us as unexplored topologies of meaning. The field of extended reality, still stumbling in the periphery of the cultural semiosphere, is becoming increasingly connected to the very central today exponential growth of artificial intelligence. This paper explores the nature of this intersection and brings forward arguments for tackling the understudied questions of its social and artistic impact. Most studies currently examine the application of AI tools within the context of XR, not the other way around. Their focus is on collecting and analyzing data regarding the speed and effectiveness of creating realistic XR worlds, the user experience in its performance-driven aspects, interactivity and gestural locomotion techniques, as well as perceptual experiments based on studies of virtual assistants. This indicates a gap of research papers addressing the topic of AI’s impact on the cultural and social mechanics of XR. This paper bases its findings and proposals on several research papers conducted on the topic of social and artistic impact. These include for example “Extended Reality and Artificial Intelligence’s Ethical Crossroads: From Sensory Manipulation to Creative Disruption” and George Mason University studies on “Automatic Generation of VR Scenarios and XR Experiences”, as well as XR productions created by leading artists and studios in the field. On the one hand, the aim is to articulate a discussion on the upcoming changes in the field of artistic practices as a result of the introduction of AI toolkits, while on the other, to contextualize the use of artificial intelligence in the realm of XR’s potentialities for new types of networking.

Humans as natural-born cyborgs: Scrutinising AI’s narrative intelligence within the 5E cognition framework

Inna Livytska — Tue, 30 Dec 2025 00:00:00 +0000

Recent advances in large language models (LLMs) and their ability to generate content align with both perspectives: AI as a powerful assistant and AI as a potential challenge to human cognition. However, these advancements expose the persistent limitations of AI compared to humans. One such limitation, as demonstrated by the folk narrative hypothesis (Hutto 2008), is the uniquely human ability to engage in storytelling, a fundamental and ancient mechanism for memory, information storage, identity formation, and world-making or world-disruption. The ability of AI to generate coherent texts is often mistaken for its intelligence, raising false claims about AI self-consciousness and sentience. This paper argues that human narrative intelligence as a drive for sense-making might turn imperfect AI text generation into a meaningful construct despite cognitive dissonance in the communication. By analysing the technical prompts used by ChatGPT-4 to generate narrative text and its understanding of key text production categories (coherence, fluency, complexity, etc.) within the 5E Cognition framework (Stilwell & Harman 2019), I will attempt to show how AI-generated narrative structures acquire full and meaningful interpretation when perceived by a human. The generative ability of AI which emerges in co-communication with a human cannot stand for its intelligence, since it reflects human perception and our intrinsic ability to hypothesise and engage with communication agents.

Needle in a needle stack: How AI causes semiotic inflation which causes experiential devaluation

Marc Champagne — Tue, 30 Dec 2025 00:00:00 +0000

While most discussions of generative AI center on issues such as algorithmic bias and disinformation, we should also consider the quantitative sea change brought about by these technologies. Large language models can generate contents at a rate uncoupled from the human datasets they were trained on. Forecasts about such artificial outputs are difficult to make, but it seems clear that the word count and image/video bank of the internet will grow far beyond what humans can actually produce. Although this growth may seem benign, I worry that it results in a semiotic inflation capable of devaluing many human experiences. What connects quantitative you discover a diamond, you become rich. When you discover two diamonds, you become twice as rich. However, when you discover a stash of diamonds so large that diamonds outnumber gravel, you become poor and instantly make every diamond owner poorer too. Similarly, AI’s vast output risks devaluing experiences that are vital to human flourishing. Adopting a wide evolutionary vantage that gives weight to proven cultural adaptations, I suggest that some situations have natural sign-to-object ratios. Hence, being flooded with too many signs can go against the long-term interests of users. Critics of AI typically cling to features which computers allegedly cannot mimic. Semiotic inflation, however, enables us to accept the possibility of perfect AI counterfeits and still detect them en masse, via their negative experiential effects. Hardcoded features like Bitcoin’s 21 million token ceiling show that “[i]n contrast to the linguistic sign, the money sign cannot be reproduced in arbitrary quantity” (Bankov 2023: 117). Prompted by the rise of generative AI, we are about to realize that non-monetary signs also cannot be reproduced in arbitrary quantity. I thus draw parallels between healthy semiotic systems and the constraints governing viable monetary systems.

Translating the wild: AI, semiotics, and the future of animal communication

Nicola Zengiaro — Tue, 30 Dec 2025 00:00:00 +0000

This article critically examines the recent rise of AI-based attempts to “translate” animal communication: with a specific focus on aquatic species such as dolphins and whales. Drawing on biosemiotics, ecosemiotics, and Umwelt theory, it argues that such projects, while technologically sophisticated, risk reducing animal semiosis to codifiable data structures and computational approximations of meaning. Through an analysis of DolphinGemma and CHAT, the article exposes the epistemological and semiotic limitations of current AI models, contrasting them with embodied, context-sensitive modes of communication in non-human species. Rather than serving as transparent translation tools, AI systems should be understood as technosemiotic infrastructures which may support new interspecies resonances, provided they are embedded within critical and relational frameworks. The analysis advocates for an ecotechnical semiotics which redefines intelligence and communication as emergent, situated, and materially grounded processes, resisting both symbolic reductionism and techno-utopianism.

ChatGPT in higher education: A semiotics investigation between cultural explosion and encyclopedic knowledge

Danilo Petrassi — Tue, 30 Dec 2025 00:00:00 +0000

Generative Artificial Intelligence, and particularly ChatGPT, today represents a crucial moment in the evolution of knowledge production and dissemination in education, echoing what Juri Lotman describes as a “cultural explosion.” As AI-generated texts flood academic and creative spaces, the question arises today is: how does this affect research methodologies and pedagogical frameworks? This paper explores the transformative role of ChatGPT in higher education, positioning it as both a disruptive force and an innovative collaborator in the knowledge-making process. Through the perspective of semiotics and Umberto Eco’s “encyclopedic model”, ChatGPT’s functioning mirrors human interpretative processes – drawing upon a vast corpora of texts, identifying patterns, and generating plausible continuations within cultural discourse. However, unlike human scholars, ChatGPT lacks intentionality, challenging traditional epistemological models which rely on authorial agency and context-dependent inference. This research investigates how ChatGPT’s generative capacities align with the rhizomatic structure of knowledge, where meaning is not linear but formed through a dynamic network of associations. However, today in higher education ChatGPT is redefining the role of teachers and learners, transforming classrooms into dialogic spaces where GenAI acts as a semiotic mediator rather than a mere tool. By facilitating inferential learning – where students engage critically with AI-generated outputs – teachers can cultivate deeper meta-cognitive awareness. In order to investigate how students engage with ChatGPT as a semiotic mediator in education, this study employs a mixed-methods approach, integrating qualitative analysis with a quantitative survey. The qualitative phase explores the perceptions of students regarding the role of AI in learning through a semiotic investigation, examining its impact on meaning- making, epistemological challenges, and pedagogical transformation. The quantitative phase consists of a survey conducted among university students across different departments (N=20). It assesses ChatGPT usage patterns, trust levels, critical evaluation behaviors, and concerns regarding misinformation and institutional regulation. The survey findings reveal that while students frequently use ChatGPT for learning and academic purposes, they exhibit a balanced approach of trust and skepticism – engaging in critical cross-checking of AI-generated content. Despite recognizing ChatGPT’s efficacy in simplifying complex topics, students do not widely use it for deep cognitive engagement or reflective academic discussions. Concerns over bias and misinformation remain significant, and while many acknowledge the need for institutional guidelines, there is also a prevailing optimism about GenAI’s future role in education. These findings suggest that ChatGPT is perceived not as a replacement for traditional learning structures, but as a tool requiring critical literacy and careful mediation. This paper argues that rather than replacing human intellectual labor, ChatGPT nowadays amplifies the cultural explosion by accelerating the translation of knowledge, making higher education both more accessible and more complex than ever before.

AI, semiosis, and the future of language acquisition: A global approach to educational semiosphere

Ilaria Ingrao, Seyedeh Maede Mirsonbol — Tue, 30 Dec 2025 00:00:00 +0000

This study explores the role of artificial intelligence in second language (L2) acquisition and the tremendous opportunities that it creates for personalized learning. As the 2030 EU agenda for quality education suggests, the emphasis of research and practice with AI-based approaches should be to foster the learning process. For this reason, the theoretical aspect of this research examines how AI can be situated in the L2 context for promoting personalized learning experiences. The study, specifically, evaluates the effectiveness of the current role of AI in L2, critically reviewing its reflection on the ethical implication of AI in education using a semiotic anal ysis. The objective is to understand how AI adapts to students’ cognitive and individual needs to be used as a methodology in learning frameworks. Additionally, the research raises two questions which address the role of AI in Barthes’s (1977) author-reader approach, proposing a semiosis-based learning model adapted from Eco’s encyclopaedia (1976, 1984) and Lotman’s concept of semiosphere (1985). This model aims to ensure deeper personalized learning by reversing traditional pedagogic approaches and analysing the role of each element (student, teacher, text, and AI) inside the educational sphere. In conclusion, the study highlights how AI-driven tools can be integrated into L2 education, facilitating communication, content creation, and engagement with new topics in an inclusive manner. Artificial Intelligence, Second Language Acquisition, Semiotic Analysis, Semiosphere, Quality Education ysis. The objective is to understand how AI adapts to students’ cognitive and individual needs to be used as a methodology in learning frameworks. Additionally, the research raises two questions which address the role of AI in Barthes’s (1977) author-reader approach, proposing a semiosis-based learning model adapted from Eco’s encyclopaedia (1976, 1984) and Lotman’s concept of semiosphere (1985). This model aims to ensure deeper personalized learning by reversing traditional pedagogic approaches and analysing the role of each element (student, teacher, text, and AI) inside the educational sphere. In conclusion, the study highlights how AI-driven tools can be integrated into L2 education, facilitating communication, content creation, and engagement with new topics in an inclusive manner.

Consumer attitudes to AI content generation in social media

Martin Varshev, Eduard Marinov — Tue, 30 Dec 2025 00:00:00 +0000

The rapid advancement of artificial intelligence (AI) has transformed content generation in social media, prompting both excitement and skepticism among consumers. This study explores consumer attitudes toward AI-generated content, drawing on a comprehensive survey of social media users. The survey, conducted across diverse demographics, examines awareness, perception, trust, and engagement with AI-generated content. The results indicate that while awareness of AI content is high, with 88.3% of respondents recognizing it in their social media feeds, attitudes towards such content remain largely negative. Many users perceive AI-generated content as “inhuman” and “boring,” with significant con cerns about authenticity and data privacy. A striking 65% of participants advocate for transparency, urging brands to disclose the use of AI in content creation. Furthermore, trust in AI content is low, with many respondents expressing reluctance to engage with or share AI-generated material, particularly when its origin is disclosed. A case study focusing on Coca-Cola’s AI-generated advertisements underscores potential brand impact, revealing that a majority of users failed to identify AI content and would prefer traditional, human-generated ads. The findings suggest that while AI has potential in content marketing, its acceptance hinges on ethical considerations, transparency, and the ability to complement rather than replace human creativity. This study offers valuable insights for marketers navigating the evolving landscape of AI in social media, highlighting the need for strategies that prioritize consumer trust and authenticity.

Use of AI in the context of fashion and related industry

Dilyana Orlova — Tue, 30 Dec 2025 00:00:00 +0000

Artificial intelligence (AI) is revolutionizing everyday tasks, enhancing processes and improving efficiency, positioning it as a powerful tool for optimizing time for producers and customers. Beyond efficiency, AI influences cultural behaviors and aesthetic preferences, opening up new opportunities for creative expression. In the fashion industry, AI supports designers and retailers by offering personalization and automation to enhance user experience amidst a landscape of abundant fashion choices. This paper examines the impact of AI on the fashion industry as both a technological innovation and a cultural influencer, referencing Gilles Lipovetsky’s theories on consumer psychology. The semiotic and dynamic nature of fashion is touched upon through the theory of Yuri Lotman (1922–1993) and Roland Barthes (1915–1980), as well as the dynamic contemporary processes and technologies through the perspective of Lev Manovich (1960–) and Emanuele Arielli (1972–). AI is shown to drive creative exploration and cultural diversity within fashion, while highlighting the human element in the aesthetic process.

The cultural explosion of AI: Navigating the intersection of artificial intelligence, society, and culture from a semiotic and interdisciplinary perspective

Kristian Bankov, Federico Biggio — Tue, 30 Dec 2025 00:00:00 +0000

This issue of Digital Age in Semiotics and Communication is dedicated to the cultural explosion of artificial intelligence. Drawing on Lotman’s notion of “explosion,” it aims to describe the “rise” of generative AI by retracing its cultural evolution and highlighting the features emphasized by the Russian semiotician: the sudden crossing of “boundaries” by extra-cultural or extra-semiotic entities, the tensions it generates between centre and periphery, conflicts among heterogeneous elements and systems, and the undeniable sense of unpredictability and destabilisation it produces. From this perspective, it focusses on the ways in which media technologies for the automatic generation of multimedia content have crossed the “laboratories” of computer science and entered the “pop” media ecologies of digital societies.

Notes for contributors

Editorial Board — Tue, 30 Dec 2025 00:00:00 +0000