AI and Communication Research:
A Conversation on Hype, Contexts, and Practices (Part 1)

RONI DANZIGER [1]

HADAR LEVY-LANDESBERG

Tel Aviv University, Israel

AYA YADLIN

Bar-Ilan University, Israel

IDO RAMATI

The Hebrew University of Jerusalem, Israel

LIDOR IVAN

Tel Aviv University, Israel

ILAN MANOR

Ben-Gurion University, Israel

On June 9, 2025, the DAN Department of Communication at Tel Aviv University convened a roundtable to explore the evolving intersections of artificial intelligence and communication research. The event brought together communication scholars from multiple universities in Israel for a wide-ranging conversation about AI’s place in the field of communication—as an object of study, a methodological tool, and, increasingly, as part of a broader cultural and technological climate that shapes new horizons for analytical thinking. The article is an edited and truncated version of the panel: “AI in Communication Research.” The panel was hosted and moderated by Dr. Hadar Levy-Landesberg and Dr. Roni Danziger, and features Dr. Aya Yadlin, Dr. Ido Ramati, Dr. Lidor Ivan, and Dr. Ilan Manor.

Keywords: artificial intelligence, communication research, media studies, trust, authenticity, interdisciplinarity

Roni Danziger:

Artificial intelligence (AI) has become a central topic in the field of communication studies—not only as an object of research but also as a methodological tool and as a cultural context that reshapes the very questions we ask. From language-based platforms to automatically generated visual imagery, AI technologies influence meaning-making processes, patterns of trust and authenticity, and the relationships between humans and technologies. This panel brings together experts in different aspects of communication studies, including, but not limited to media studies, sound and voice studies, visual communication, ethics, Internet studies, discourse analysis, the history and philosophy of media, and digital diplomacy. The experts were asked key questions, discussed below, to explore the place of AI in communication research as well as the challenges and opportunities it presents for scholars in the field. This is the first part of the roundtable, which focused on foundational research approaches, conceptual roles of AI, temporal perspectives, and cross-disciplinary engagement.

First Encounter With AI: Natural Continuation or a Shift in Scholarly Practice?

Hadar Levy-Landesberg:

Our panel is a roundtable format. We will begin with questions about each of your personal research contexts, then broaden the discussion to theoretical, methodological, and ethical issues. Our first question is: In your experience or interest in artificial intelligence within research, when and how did your encounter with the field begin? Was it a natural continuation of your previous work, or a complete shift into a new area?

Aya Yadlin:

For me, it was a natural extension of my previous work analyzing news coverage—specifically, the narratives journalists use when presenting new AI technologies to the public. My first encounter was in 2020 with the emergence of deepfakes. This built on earlier research I had done on framing and narrative construction but shifted toward new, “easy-to-use” technologies accessible to nonexpert users. Another dimension that expanded my interest was the global scope: Previously I focused mostly on Israeli coverage of global and local issues, but here we conducted a cross-national analysis of global reporting on the phenomenon.

It was an experience that shifted the focus from expert discourse to general public engagement. It was one of the first times I studied a phenomenon before it became widely popular. In those early days, the discourse was very “pure”—we didn’t yet fully understand what deepfakes were, nor how to connect them to other contexts. That made them easier to analyze as a distinct phenomenon, similar to other emerging technologies.

Ido Ramati:

Thinking back, I can pinpoint three key moments. First, in 2016, while writing an article about Google Translate with Amit Pinchevski, Google suddenly shifted to a more advanced model, forcing us to reconsider the platform’s history. We realized the history of Google Translate could be told through the broader history of AI, dating back to the late 1940s and early 1950s—the era when the term artificial intelligence first emerged. Second, during my postdoc at the International Research Institute for Cultural Technologies and Media Philosophy (IKKM) at Bauhaus-University, Weimar, the initial GPT model—still text-only—was released. I sensed a major disruption but did not fully understand how it worked. This resonated with my long-standing interest in the intersection of language and technology. Third, after ChatGPT integrated multiple models, its societal impact suddenly became undeniable. This marked a shift into somewhat new territory for me, while still building on my earlier focus on language and technology.

Lidor Ivan:

My first encounter with AI was, honestly, methodological convenience. For my postdoc research on perceptions of authenticity, trust, and deception in online dating profiles, I needed dozens of manipulated images. Traditionally, this would require multiple photoshoots and extensive editing. Around that time, I learned about Midjourney’s text-to-image capabilities and decided to experiment. I generated about sixty images and ran a pretest. Two surprises emerged: First, most participants perceived the AI-generated images as real photographs, and second, even when they recognized them as AI-generated, this did not significantly affect trust or perceptions of the depicted person. This contradicted my initial assumptions and convinced me AI imagery was not just a methodological tool but a social phenomenon worth deeper investigation. This experience marked the moment I realized that AI-generated imagery was not merely a convenient tool for stimulus creation but a phenomenon in its own right—one that raises new questions about how visual content is interpreted, trusted, and socially integrated.

Ilan Manor:

My studies examine the impact that digital technologies have on diplomacy and on diplomatic actors, which includes diplomats, foreign ministries, and national leaders as well as terror groups. At the moment, foreign ministries across the world are examining how they can best integrate AI into their daily work. This includes everything from using AI to draft speeches and analyze press coverage, using bots to converse with digital publics and promote foreign policies, and developing in-house AI tools. These can help MFAs search vast digital databases and enhance decision making capabilities. For instance, instead of ChatGPT, the U.S. State Department is hoping to build a State GPT that will be of use to American diplomats. For instance, American diplomats may use such a tool to search for historic patterns in the negotiating tactics of different states and best prepare for new negotiations. However, AI also poses a real challenge to diplomacy, be it in the use of deep fake AIs for disinformation or the use of AI bots to automate the circulation of fake news and disinformation. I have been asked by various MFAs such as the State Department and the EU to assess how AI could be best managed internally and what effective AI regulation would look like.

AI: Object of Study, Methodological Tool or Cultural Context?

Hadar Levy-Landesberg:

This brings us to the next question. AI can take on many roles in research: It can be the object of study, a methodological tool to aid our work, or the broader context in which we operate—shaping the questions we ask and how we understand media and interaction with technology. How would you define the place of artificial intelligence in your own research? How does this connect to your overall research approach?

Lidor Ivan:

I do not study the technology itself. My focus is on the place that the outputs of the technology—rather than the algorithms—occupy within the discourse on authenticity and trust in interpersonal, and often intimate, interactions such as romantic relationships. My aim is to examine societal reactions to these outputs (i.e., images), how they vary across cultures and over time, and whether they are considered legitimate to use when presenting oneself. In this sense, my work continues earlier research on visual “manipulations”—I use that term in a technical sense, not a moral one—performed on images in interpersonal contexts. AI-generated imagery is, for me, another point on the same spectrum. It may differ technically from traditional photo editing, but what interests me is how it differs socially: whether and when it is considered acceptable or unacceptable to use such imagery in self-presentation and persuasion.

Ido Ramati:

Unsurprisingly, my answer is “all of the above.” AI is an object of study: I examine the technology itself, its development, and its conceptual implications—particularly in my current focus on how the human voice is perceived in the AI era. But AI is also a tool. For example, at a recent conference in Chicago, part of my presentation involved a live “conversation” between me and an AI voice model. I was not the only one doing this—the entire panel incorporated recorded voices—offering a kind of performance that invited reflection on what AI can contribute both as an object of study and as a tool. For me, these categories—tool, object, and conceptual framework—are not separate. I see value in holding them together. This approach allows me to remain critically aware of myself as a researcher, the tools I use, and the objects I study.

This also connects to the historical perspective that underpins my work: As media scholars, we give ourselves permission to study phenomena as they unfold, even though historians might insist you need 70 or 90 years before calling something “history.” But we also have a responsibility not to simply follow the hype, especially since hype is often driven by political, social, and industrial forces promoting these technologies. Sometimes we need to take a few steps back, even if we don’t fully understand what an algorithm is doing to us in terms of voice, text, or global context. That distance is crucial for critical and responsible scholarship.

Aya Yadlin:

For me, it is also “all of the above”—and then some. On the surface level, AI is a data-gathering tool, which immediately raises ethical questions about what should or should not be included, whether data collection should occur in closed environments or be left to roam freely online.

As an object of study, I do not examine AI outputs in the same way Lidor Ivan does. My focus is on AI as a representation of broader structures of power and societal value systems. For example, my studies of deepfakes and ChatGPT—conducted four years apart—look at how journalists talk about these technologies. But beyond the technologies themselves, the coverage reveals the society’s techno-moral framework.

And as an editor of an academic journal, I also face practical questions: How should we handle submissions where reviewers suspect ChatGPT contributed to the writing? What expectations should we set for reviewers? These considerations have shaken my epistemological assumptions about what counts as “knowledge” produced by researchers. If we use AI for analysis, what are we actually hearing in the research output? Whose “voice” is it?

This indicates a deeper cultural shift in academia: While recent years have marked a move away from the dominance of the White, Western male researcher toward more diverse and reflexive perspectives, we are now confronted with the reality that the designers of AI platforms—and their biases—are shaping the outputs. Personally, this has pushed me back toward manual textual analysis as a reminder of the importance of close, human-driven interpretation.

Hadar Levy-Landesberg:

And yet, as you and others have noted, AI can also function as a “member” of the research team—providing tools for reflexivity and reliability in qualitative research, even if we can’t always know what stands behind its outputs.

Roni Danziger:

Exactly—and that blurs the boundaries of authorship. Are we co-authors with AI? If AI cannot be held responsible for what it produces, and its creators also avoid responsibility, where does that leave accountability?

Ilan Manor:

Following up on Aya’s comments about journalistic coverage of AI systems and their covert biases, I am currently conducting several studies that all deal with AI and diplomacy. One field that interests me is understanding how generative AI tools impact our perceptions and beliefs about nations. I think that AIs have a potentially dramatic impact on our beliefs and worldviews because of the “mystification of AI” or the way AIs are framed in the media. Media reports tend to suggest that AIs are incredibly accurate and sophisticated, so sophisticated that they have passed entry exams to Harvard Business School, Ivy League Universities, and Medical Licensing Exams. This might translate into public trust of AIs. But these are inherently biased. Two studies conducted with Elad Segev asked ChatGPT to say 10 good things and 10 bad things about different countries. We clearly saw a pro-Western bias and a clear bias between countries in the Global North and Global South. Global North countries were depicted functional states with robust economies and great places to visit and invest in. Global South countries were all depicted as suffering from endemic corruption and violence and as terrible places to visit or work in. In “bad things” about France, ChatGPT listed warm summers and long lines to see museums. In the case of South Africa, ChatGPT mentioned poverty, corruption, failing medical systems, inadequate education, and gun crimes. This is important as negative depictions on AIs may shape how people view states and regions. This is a crucial finding for diplomats who now conduct massive digital campaigns to manage their national image and brand. Effective nation branding may now rest on countering negative depictions on AIs.

I also examine how AI’s country of origin shapes answers about global affairs. For example, U.S. AIs, European Ais, and Chinese AIs offer very different answers to questions like, Why does the United States support Israel, or Why does the United States support Ukraine in the war against Russia? This is really fascinating to me, as clear ideological differences emerge in the replies to prompts.

AI in Temporal Perspective: Past, Present, and Future

Hadar Levy-Landesberg:

In addition to differences in research questions, methods, and objects of study, you also approach AI from distinct temporal perspectives. Ido Ramati works historically, tracing the antecedents of phenomena we identify today; Aya Yadlin and Lidor Ivan focus on contemporary developments and discourse; and Ilan Manor explores possible future scenarios for AI in diplomacy. How does your chosen time frame influence the way you study AI? What advantages and challenges come with working from that perspective?

Ido Ramati:

Speaking as the “past” representative—when Aya Yadlin raised those fundamental epistemological questions earlier, I was reminded of Matthew Kirschenbaum’s (2016) book on the history of word processors. He shows that similar dilemmas arose in the late 1970s and early 1980s: If a computer’s spellchecker corrects my text, does that say something about my linguistic abilities? Is the text still “mine”?

Stepping back, I see this as part of a broader, transhistorical category: human–machine relations. They have existed throughout human history, from the earliest tools onward. The advantage of such historical perspective—or what Siegfried Zielinski (2006) calls the “deep time of media”—is that it tempers the panic that often accompanies technological change. It allows us to look at AI with a cooler, more reflexive eye. That said, the challenge is to avoid the trap of claiming “nothing is new.” Yes, there are continuities, but there are also real novelties. The task is to walk that line—acknowledging precedents while identifying what is genuinely different. That interplay is what makes it so intellectually rewarding.

Lidor Ivan:

My work is firmly rooted in the present—though, as I mentioned earlier, that happened somewhat by accident. I did not expect the results I found in my pretests or main study. The findings point to shifts in how people relate to visual imagery and photographs that I believe are unfolding right now. The advantage of studying the present is that I can document these changes as they happen, perhaps even influencing our understanding while the processes are still forming. The disadvantage is that the ground is unstable. We do not yet have a solid theoretical or empirical base from which to draw firm conclusions. There’s a risk of capturing only a snapshot without fully understanding the longer-term implications.

Aya Yadlin:

I also study the present, but I’ll touch a bit on the future. The main advantage is that I do not need to reinvent the wheel regarding the broader discursive frameworks: The well-established “utopia vs. dystopia” narratives about technology are already there, and I can insert new technologies like AI into those existing debates. But I share Lidor’s sense of instability. The digital literacy required to critically engage with AI—and to teach about it—is still developing. Without established regulations or ethical norms, we find ourselves in a constantly shifting frontier. I have had moments of realizing that something I taught my students last week was already outdated because of a development I learned about yesterday. This creates a kind of identity crisis as a researcher and educator: Are we producing the best possible scholarship if we cannot keep pace with the changes? I believe that in five years, we will have far greater literacy around these issues—but, ironically, we cannot know what that literacy will look like until we get there. It is an endless chase.

Ilan Manor:

One of the crucial issues regarding AI in my opinion is that it creates temporal chaos. In my world, diplomats are eager to use AI for forecasting, simulations, and real-time crisis management. For instance, simulating a Russian invasion of Lithuania and then simulating different responses by NATO while trying to identify which response would yield the best outcome—a Russian withdrawal and peace. The main problem here is that diplomats make an inherent assumption that the future will mirror the past. In fact, all AI is based on that assumption. AI is always trained on past data. It is familiar with past events, past crises, past actions. Using AI to predict or forecast the future thus creates a temporal paradox where the past is superimposed in the future. But by nature, crises are marked by novel features and two crises are identical. We can now use AI to prepare for another COVID-19. But the next great pandemic will certainly differ from COVID in various ways and so all our use of AI will be useless. Moreover, by using AI we apply past remedies to future problems, creating a temporal loop. What we need is new ideas and new approaches, not the same old tactics that have proven limited. AI limits this kind of “new” or novel thinking, as it is limited to the past.

AI and Interdisciplinarity

Hadar Levy-Landesberg:

Our discipline is characterized by a wide range of theoretical and methodological approaches. Has working on or with AI led to new interdisciplinary connections across fields in your work, or in the work of colleagues in your field? Has it required you to step out of your comfort zone and acquire new languages—technological, methodological, or theoretical—in order to carry out your research?

Lidor Ivan:

Based on prior literature on visual imagery and the so-called truth value of photographs, I expected AI-generated images—depictions of things that do not actually exist—to be perceived as inauthentic, untrustworthy, even intentionally deceptive. This was consistent with earlier research on edited photographs. But in my study, that expectation did not hold—at least not yet. I did not find a significant difference in trust between AI-generated and real images. This was surprising, but it did not necessarily push me to venture outside the communication field. Our field already integrates concepts from technology studies, such as the social construction of technology. I did have to deepen my engagement with ideas that were not central to my prior work, but it still felt like working within the broad parameters of communication studies.

Ido Ramati:

In my case, I thought interdisciplinarity would come from collaborating with computer scientists, but that has not happened yet. The most surprising connection was with theater studies at Tel Aviv University. I have been collaborating with my colleague Ruthie Abeliovich in the Department of Theatre. We have co-authored two articles, and I have learned a great deal from her perspective—seeing AI not just as a “social actor” in the sociological sense but literally as a performer. For example, we wrote about Yiddish-language TikTok and the role of algorithmic voices there, and also about the history of sound recording in theater—how technologies like the phonograph paved the way for the AI era. This is a very long historical arc. AI was not the direct cause of our collaboration, but it sparked a curiosity in me for looking beyond the traditional boundaries of communication and finding conceptual overlaps in other disciplines.

Aya Yadlin:

I have thought about this question a lot. Since communication studies is already inherently interdisciplinary, I did not find myself forging entirely new theoretical frameworks for my research. But in teaching—which is inseparable from research for me—I have had significant cross-campus collaborations. One example is a project with colleagues in history, law, and education on using ChatGPT as a “teaching assistant” within a closed system prompt. The AI cannot fetch information from the Internet; we’ve run a two-year pilot integrating it into our courses. I was invited as a “media expert,” but I quickly realized others saw communication scholars as central to shaping the critical perspective on this technology—sometimes more so than, say, education scholars. This was affirming for our field, but it also made me question whether we always know best. Why not let sociologists or others lead in certain contexts? This has opened valuable interdisciplinary conversations, especially in pedagogy, even if less so in my own research.

Roni Danziger:

A surprising collaboration for me, though not “interdisciplinary” in the formal sense—since we are both communication researchers—is my current project with Hadar Levy-Landesberg about interactions with AI voice companions. I am a discourse analyst, and Hadar researches voice, sound and media. I think AI has a special capacity to spark such partnerships, whether within or across disciplines, because it affects so many areas at once.

Ido Ramati:

Exactly—and that is the point. AI is not one thing. The term “AI” has become almost a cliché; it is neither truly “artificial” nor necessarily “intelligent.” Under the AI umbrella, we have a vast diversity of models—natural language processing, machine learning—that long predate the current hype. That breadth means AI is not just transforming one field; it is impacting many at once. As a result, collaborations arise in multiple places simultaneously.

Aya Yadlin:

And it is also about the utility of AI. Unlike, say, television representation research—which remains largely within media studies—AI is now used by everyone, not just experts or media producers. This broad adoption creates more opportunities for cross-domain collaboration, simply because it is part of everyday practice for so many people.

Ilan Manor:

My research has become much more diverse due to AI and thanks to AI. I now work with colleagues from the fields of semiotics, computational propaganda, machine learning, network analysis, and sentiment analysis. My research is much more diverse but also constantly evolving as new AI tools and new AI applications are invented. The pace of AI innovation is relentless, and sometimes I really struggle to keep up.

Conclusion

This first part of the roundtable examined how AI enters communication research as an object of study, a methodological tool, and a cultural context, while reframing debates about media histories, academic practices, and disciplinary boundaries. What emerges is a view of AI as a contested phenomenon within communication scholarship—simultaneously something we study in the world and something that acts upon us as researchers, educators, reviewers, and editors. Through the roundtable format, AI appears not as a fixed object but a dynamic site where pedagogical practices, methodological choices, and scholarly infrastructures are negotiated alongside broader technological, political, and cultural struggles. Part two, in the same Forum, will extend this conversation to questions of trust, authenticity, and the future of communication scholarship amid the transformations brought by AI. Together, the two parts call for critical and interdisciplinary approaches that recognize the broader cultural and political stakes of making meaning with machines. At a moment when AI is reshaping the very questions communication scholars ask, the roundtable format underscores the importance of intradisciplinary dialogue.

References

Kirschenbaum, M. G. (2016). Track changes: A literary history of word processing. Cambridge, MA: Harvard University Press.

Zielinski. S. (2006). Deep time of the media toward an archaeology of hearing and seeing by technical means. Cambridge, MA: MIT Press.

Biographies

Dr. Aya Yadlin is a senior lecturer at Bar-Ilan University’s School of Communication. Her work explores the intersection of culture, politics, and digital media, particularly processes of identity construction and cultural debate in online platforms. She also studies ethical aspects of digital communication, including privacy, social responsibility, and information ethics.

Dr. Ido Ramati is a senior lecturer in the Noah Mozes Department of Communication and Journalism and in the Program in Cultural Studies at the Hebrew University of Jerusalem. His book Lingua Ex Machina (University of Pennsylvania Press) examines media technologies’ historical role in the renewal of the Hebrew language. His recent articles address the status of the human voice in the AI era, media ventriloquism, and the theatricality of voice-based AI agents such as Siri and Alexa.

Dr. Lidor Ivan is a visual communication scholar and postdoctoral fellow at Tel Aviv University, specializing in the study of persuasive uses of visual imagery in interpersonal and mediated contexts. His research examines how visual manipulations, including AI-generated images, influence perceptions of trust, authenticity, and deception, with a broader interest in the role of visual media in shaping social evaluation.

Dr. Ilan Manor is a senior lecturer at Ben-Gurion University’s Department of Communication, specializing in digital diplomacy and public diplomacy in the age of social media. He examines AI’s impact on public narratives, opinion formation, and international messaging strategies.

Dr. Roni Danziger is a senior lecturer in the DAN Department of Communication at Tel Aviv University. She studies language and social interaction (LSI), with a focus on interpretation and meaning-making processes. Her research interests include sociopragmatics, politeness, discourse analysis, language in social media, intercultural communication, and positive communication.

Dr. Hadar Levy-Landsberg is a senior lecturer in the DAN Department of Communication at Tel Aviv University. Her research explores the intersections between voice and artificial intelligence technologies and their social and cultural implications, with an emphasis on historical perspectives. Her areas of interest include voice and sound studies, media theory and philosophy, history of technology, and critical data studies.

[1] This publication was supported by the Alon Fellowship.

AI and Communication Research: A Conversation on Hype, Contexts, and Practices (Part 1)

AI and Communication Research:
A Conversation on Hype, Contexts, and Practices (Part 1)