Ask Me Anything ◕‿◕ - How ChatGPT Got Hyped Into Being

JASCHA BAREIS

University of Fribourg, Switzerland

This study reconstructs how chatbots like ChatGPT got hyped into being. It dissects the actors and dynamics that triggered, fueled, and disseminated the hype. Through the lens of hype studies, the article interrogates three empirical realms: (1) company websites where the chatbots are presented, (2) blog entries and newspaper interviews by prominent tech figures from Silicon Valley, and (3) New York Times articles. The study shows how chatbot hype is driven by a dynamic between privileged actors (hypers) and a media frenzy, both influencing and being carried by the public and politics alike. The following different interdependent building blocks in the chatbot hype are identified: (1) strategic ignorance—depicting large language model chatbots as knowledge models, (2) the weird and eerie—panicking about the uncanny side of chatbots, (3) the battle—staging a spectacle of competition among tech giants, and (4) crossing the line of the normal—praising the dualism of an apocalypse or a tech-religious calling. The article unravels the core circulated narrative that turns the hype into a powerful societal phenomenon.

Keywords: ChatGPT, chatbot, hype, big tech, sociology of expectations, large language models, Xrisk, catastrophic risk, hallucination, spectacle, Silicon Valley

Jascha Bareis: [email protected]

Date submitted: 2024-09-24

In February 2023, Sam Altman, the CEO of OpenAI, which had just released the chatbot ChatGPT, declared in the mission statement “Planning for AGI [Artificial General Intelligence] and Beyond,” “Successfully transitioning to a world with superintelligence is perhaps the most important—and hopeful, and scary—project in human history” (Altman, 2023, p. 26). The quote captures well the critical historical momentum, ChatGPT represented for many tech enthusiasts, hailing it as a milestone toward a unifying destiny of holy machinery superintelligence. No matter the plausibility of this trajectory, the release of the chatbot caused a societal frenzy across media, politics, and user bases. ChatGPT added 100 million users by February 2023, making it the fastest-growing application ever (Thubron, 2023).

How is it possible that the release of a chatbot can trigger callings for transcendental futures and evoke such ecstasy? This study conceptualizes this societal phenomenon as hype, arguing that the overpromising of chatbots entertains a different mode of futuring than just proclaiming and inviting a vision. Instead of cultivating future literacy, the chatbot hype is driven by an entertaining media spectacle and by the opportunism of some reckless private actors. Hype is so far rather poorly carved out as a scientific and analytical concept, colloquially being treated as some mere folk talk or marketing prose.[1] However, I argue the concept of hype can help answer the following research questions: What triggered the social frenzy around chatbots? What fueled and disseminated it? To investigate the case study of the chatbot hype, the article interrogates three societal realms: first, the presentation from the company websites where chatbots are hosted; second, commentaries on X (formerly Twitter) and blog entries by prominent tech figures who enjoy a high degree of public authority and a prominent speaking position in media, given their alleged expertise in the field of AI or perceived “visionary” achievements; and, third, New York Times newspaper articles in the time span between November 2022 and August 2024, by journalists covering chatbots and interviewing prominent tech figures about their impact.

The fact that technology can trigger expectations and dreams of exuberant achievements has been investigated by many contributors to cultural and media studies as well as science and technology studies (STS). Dynamics of expectations through rhetoric, stage, and imagery dramatization have been studied with the presentation of new tech gadgets like the iPhone (Sharma & Grant, 2011) or with showcase videos of air mobility and their projective narratives of futurist air taxis (Woznica, 2023). Similarly, the triggering dynamic of “hot air” tech talk in start-ups about potentials, missed opportunities, or failures concerning other tech company cases like Nokia, Google, or Apple has been associated with hype (Hockenhull & Cohn, 2021)—and the innovation culture of Silicon Valley more generally (Bourne, 2024; Duff, 2016).

Moving concretely to the case study of artificial Intelligence (AI), recent work has looked at AI hype in health care (Strange, 2024) and cryptocurrencies (Castro & Belsunces, 2025); the bold promises fueling and then failing the expectations of digital twins (Michalec, 2025); the narrative construction of AlphaGo as a mystical, sacred artifact (Binder, 2022); and systemized discursive typologies around AI hype (Kotliar, 2025). This work discloses AI as a phenomenon, not just as an object, that evokes emotions and performativity or is even attributed with a seductive, enchanting aura. Pointing to this aspect, historical studies on AI have discussed how the idea of outperforming and transcending the human is entertained by aspects of tech development determinism and machine opacity (Campolo & Crawford, 2020), feeding AI hype beyond simplified narratives of AI summers and winters (Galanos, 2023). Further, it has been shown that AI perception constantly mediates at the binary of hopes and fears, or redemption or doom, most concretely embodied in fictional narratives around AI (Bareis & Bächle, 2025; Binder, 2022; Cave & Dihal, 2019). Surprisingly, these narrative accounts of AI stem not from an uninformed public, but from the heart of the AI expert community from the early 1950s (Natale & Ballatore, 2020), just to be echoed in public arenas shaping overall AI sensemaking. Furthermore, they have been embraced by nation-states, triggering technological races between geopolitical competitors (Cave & ÓhÉigeartaigh, 2018), who have been urging policymakers to fire bold tech futures (Bareis & Katzenbach, 2021). In a turn toward “communicative AI” (Hepp et al., 2023), scholars in the field of communication and media studies call for distinct analysis of how the automation of communication through algorithms is “embedded within digital infrastructures” and “entangled with human practices” (Hepp et al., 2023, p. 28). This study ties in with this research agenda, but also adds a special focus on the actors, their agency, and their intentionality in starting and steering the hype dynamic around chatbots.

The article shows that many of the rhetorical elements in the aforementioned studies serve as building blocks in the construction of hype around AI chatbots. It analyzes how they are integrated into a consistent and coherent narrative building on the following four pillars, adding to the research that shows “ChatGPT isn’t magic” (Leaver & Srdarov, 2023), but a well-orchestrated social phenomenon: (1) strategic ignorance—depicting large language model (LLM) chatbots as knowledge models, (2) the weird and eerie—panicking about the uncanny side of chatbots, (3) the battle—staging a spectacle of competition among tech giants, and (4) crossing the line of the normal—praising the dualism of an apocalypse or a tech-religious calling.

The publication of this article falls inside a hype dynamic that seems ongoing. While the peak of the chatbot hype seems to have passed, it is clear the discursive phenomenon of hype has already materialized in policy outputs and path dependencies—and, as some argue, faces the threat of an economic bubble that is about to burst as “air begins to leak out the overinflated AI bubble” (Hiltzik, 2025, p. 1; see discussion in conclusion). Before beginning the analysis, the study introduces the conceptual framework of hype and then presents the empirical realms and the analytical approach in more detail.

Theoretical Takes on Hype(rs)

For long, hype studies as an established and structured research field did not exist, but has been dwelling in academia as siloed work focusing on the concept (Milne, 2020; Palavicino & Konrad, 2015; Seidensticker, 2006), with recent efforts to be carved out as a structured field with a research agenda (Bareis, Roßmann, & Bordignon, 2023; Ramiller, 2006).[2]

Especially with regard to technology and innovation trajectories, STS, technology assessment, and responsible research and innovation have reacted with future-directed heuristics to deconstruct and counter what is also coined as “overpromising,” such as the sociology of expectations (Borup, Brown, Konrad, & Van Lente, 2006), vision assessment (Grin & Grunwald, 2000), anticipation (Alvial-Palavicino, 2015), sociotechnical imaginaries (Mager & Katzenbach, 2021), or forecasting (Martino, 2003). More historical accounts trace how hype changed its meaning conceptually from an association and marker of sub- and counterculture to a more recent entrepreneurial and innovation spirit (Wadhwani & Lubinski, 2025).

Communication studies discusses and measures hypes in media circles, especially focusing on waves of media attention, also understood as “buzz,” “spin,” or “trend” in the world of news, social media, or marketing. These fields of inquiry are elaborated on more conceptually (see, e.g., Davis, 2013; or Wernick, 1991, for marketing, advertising, and promotional culture; Thompson, 2010, for publishing; and Vasterman, 2005, for news circles), but also approached empirically, for example, through studying dynamics of media coverage (Kari, Lehtonen, Litmanen, & Kojo, 2023; Wien & Emelund-Præstekæker, 2009). Quantitative approaches to measure waves of attention are also taken up by the newly emerging field of scientometrics, studying how science is influenced by writing style, metrics, and citation scores, looking, for instance, at the role of hedging words or phrasing in abstracts to attract readers (Bordignon, Ermakova, & Noel, 2021; Hohmann, Barnett, King, & Connell, 2025).

Conceptually, Palavicino and Konrad (2015) offer some typologies of hype, differentiating among hype as an analytical tool to read future market developments or map current media attention, hype as exaggeration following strategic activities of actors, or hype as a societal phenomenon looking at the discursive dynamics of expectations in society. To distinguish hype from the other future-related concepts mentioned above, this work especially includes the latter two perspectives in the analysis, stressing that hypes are both a societal dynamic and are also triggered and driven by strategic agents (hypers), pointing to the normative nature of hype. Hypers must be understood as rather problematic appropriators of futures, following a deliberate mode of future capture (Geiger & Gross, 2017). Instead of cultivating future literacy in the form of reflection, inclusion, and alternative pathways, hypers exploit the epistemic uncertainty of the future for opportunist purposes (Bareis et al., 2023). These actors instrumentalize promises for the sake of creating attention, followership, and investments. The consequent societal dynamic of hype narrows down temporal dynamics (opportunity) in the anticipation of a future, pointing to the normative dimension of hypes, as they can mobilize attention, but also relinquish democratic zones of imagination, speech, and contestation.

A comprehensive definition of hype can be found from the communication scholar Powers (2012). She defines hypes as “a state of anticipation generated through the circulation of promotion, resulting in a crisis of value” (Powers, 2012, p. 863). I take this definition as a reference point to extend toward an updated definition of hype, especially focusing on hype’s inherent normativity through the stressing of exaggerated promises and the managing of impressions and opportunity costs through emotions.

Normativity: Hype as Exaggerated Promises

In his analysis of human enhancement discourses (many of the same actors that entertain the chatbot hype), Nordmann (2007) warned of the seduction of what he called “if and then” syndrome. By means of a “radical foreshortening of the conditional” (Nordmann, 2007, p. 32), such speculative projection could create forceful and unchecked futures. This standpoint stresses the epistemological dimension when tackling the pitfalls and shortcomings of bold future proclamations, but critiques in the form of evidence, rationality, or implausibility do not capture the phenomenon of hype—to the contrary, as these aspects are rather ignored (or, worse, instrumentalized) by hypers. In contrast to Powers (2012), who stresses in her definition the resulting “loss of value” (p. 863) in hypes, I argue that someone always profits from hypes (also when a wave wanes). After all, money changes pockets; it never gets burned. Hypers are masters in coining uncertainty and vagueness about the future into an overly optimistic story of opportunity, thereby possibly misleading decision makers or overshadowing alternative futures.[3] It is not that hypers lie—it is rather that they do not care about categories of truth or lie, about the difference of fact and belief. In this respect, they also differ from pioneer communities, which have an interest in structuring and curating the future (Hepp, 2024) rather than exploiting it. Hence, epistemological categories do not grasp the phenomenon of hype. In this sense, the hyper rather makes use of what philosopher Frankfurt (2005) discusses with the concept of “bullshit.” He conceptualizes the relation between truth and bullshit in the following manner:

The fact about himself that the bullshitter hides [. . .] is that the truth- values of his statements are of no central interest to him [. . .] This does not mean that his speech is anarchically impulsive, but that the motive guiding and controlling it is unconcerned with how the things about which he speaks truly are. (Frankfurt, 2005, p. 55)

Hypers need to emotionally overstress the value of a future instead of the future being able to speak for itself (by plausibility or social appeal), making it an affective tool in promotional culture (Bourne, 2024). This again, though, shows the cleverness and wickedness of hypers. They do not care about truth, but they surely know how to use sensationalism to their benefit, or again, how Frankfurt (2005) puts it, “In comparison to liars, bullshitters are ‘more expansive and independent, with more spacious opportunities for improvisation, color, and imaginative play’” (p. 53).

Managing Impressions and Opportunity Costs Through Emotions

This aspect leads to another core pillar of hype being a social phenomenon. A hyper alone cannot cause a hype: It needs an audience to rise and be carried. This audience is not just a passive spectator, but a complicit collaborator acting like a catalyst. It is the media frenzy, the social media bursting of feeds, disseminated by the many users and journalists that give traction to hype. As Powers (2012) rightly points out, promotion needs to be circulated. That is why hypes are not only descriptive, but highly performative (action-ascribing, enabling, fueling, etc.). They create topicality; coin leaders; produce feeds; attract investors; skyrocket stocks; and give birth to start-ups, analysts, influencers, tech gurus, and other actors who know how to exploit the hype to their advantage.

However, an audience of followers not only wants to be entertained but to be impressed, in order to be converted into collaborators. As Stilgoe (2020) writes, “Technological hype is not just exaggeration, nor is it idle speculation; it is an act of persuasion” (p. 40). The repertoire of impression management in hypes involves figurative language that can trigger emotions to craft a community, such as bold statements and exaggerated claims that hail future benefits involving fabulous potentials and shiny prospects that need to be mobilized. Hypes often maneuver within the binaries of hopes versus fears, vast possibilities versus a closing window of opportunity, or redemption versus doom. A missed chance, or a bleak future if not acted on now, is a common rhetorical motive to spur action. All too often, hypes normatively set the agenda with “wishful worries,” which Brock (2019) calls “problems that it would be nice to have, in contrast to the actual agonies of the present” (p. 1). Here, especially references to history by leaders of hypes are very common, stressing the kairos moment of a small window of opportunity (Bareis et al., 2023). The promising invitation of conquering the seemingly unachievable, or visiting the never—before—visited (Beckert, 2016), can unify an audience through feelings of belonging and identity with a peer group. This social momentum also points to a psychological dimension with hype. There lies a deep wish within hypes to escape (even only for a moment) everything that entangles and complicates the innocent declaration of a bold promise or a golden future at the horizon. This escapist notion shows parallels to Roland Barthes’ (1992) analysis of the power of myths, which serve to conceal and cope with social contradictions and help citizens escape a sometimes dull and mundane routine of everyday life.

The afore elaborated leads me to the following definition of hype that extends Powers’ (2012) definition and informs the case study of chatbots with the aspects of impression management and the normativity of opportunism: Hype is a social dynamic created through emotional staging and circulation of a distinct and commendable—albeit opportunistically exaggerated and implausible—value development.

Empirical Realms of Analysis

Big tech’s and media’s framing of the chatbot phenomenon powerfully demonstrates how speaking position in the public communication arena and impression management for the creation of followership influences the acceptance of chatbots (or any other novel technology). Nonspecialist users are highly influenced by these framings. To analyze the hype around chatbots, the study investigates three empirical realms. First, the descriptions and presentations from company websites are retrieved, where chatbots are presented and hosted. The analysis covers the six most popular ones, stemming all from the Global North: ChatGPT (OpenAI), Co-pilot—before Bing Search—(Microsoft), Gemini (Google), Meta AI (Meta), WatsonX (IBM), and Le Chat (Mistral).

Special focus lies on ChatGPT, being the chatbot that unleashed and continues to nurture the hype. I use the online Internet Archive with its Wayback Machine to retrieve earlier versions of the websites to understand how the corporate depictions of the chatbots changed throughout the time span between November 2022 and August 2024 (Figure 1 and 2).

The second empirical realm includes commentaries on X (formerly known as Twitter) and blog entries and statements by prominent tech figures. These figures enjoy a high degree of public authority and a prominent speaking position in media, given their alleged expertise in the field or perceived visionary achievements. Hence, their evaluation and commentary about chatbots influence users’ trust relationship to technology and its performance. The selection of persona is not exhaustive, but resembles the most invested and influential figures, arriving at the sample by a network sampling approach. Most of the actors are CEOs in Silicon Valley or tech thinkers/funders/investors/philosophers who tend to refer to each other’s achievements (as rivals, visionaries, or companions), unmasking an elite network that outs itself through self-referencing. Notable characteristics of the traced-down sample network are as follow: They are all men (many stem from, or work in, the Anglo-Saxon world, with some exceptions given several American Indian CEOs); got their education at elite universities like Stanford, MIT, or Oxford; and are firm believers in technological progress or more extreme tech accelerationism. Most belong to the top percentage of the richest people worldwide; are advocates of (authoritarian) market libertarianism; and lead, or have a role in, the largest companies around the globe (given market capitalization). The companies, donating foundations, and worldviews of many of these prominent figures are closely intertwined. For example, Elon Musk or the Open Philanthropy organization finance many projects and start-ups featuring on the list (see Table 1). Each of these tech figures prefers a different online medium for reaching out to the public, which was acknowledged for the empirical analysis. For example, while Elon Musk uses X as his main communication channel, Bill Gates publishes his thoughts on his own blog (Gatesnotes), while Sam Altman uses the blog of the company webpage of OpenAI to preach his gospel of AGI.

Table 1. List. Prominent Tech Figures Commenting on Chatbots.

Name	Role
Gates, B.	Former CEO of Microsoft; currently investor and manager of the Bill & Melinda Gates Foundation.
Nadella, S.	CEO of Microsoft (since 2014). Microsoft is the primary funder of OpenAI.
Musk, E.	Entrepreneur and owner of SpaceX, Tesla, and X (formerly Twitter). Known for his role in the tech industry and advocacy for AI and space exploration.
Altman, S.	Investor and software developer; CEO of OpenAI. Advocates for the development of AGI.
Pichai, S.	Former consultant and entrepreneur; CEO of Alphabet Inc. (Google).
Cook, T.	Industrial engineer and manager; CEO of Apple Inc.
Zuckerberg, M.	Technology entrepreneur; CEO of Meta Platforms, Inc.; founder of the social media platform Facebook.
Wozniak, S.	Technology entrepreneur and computer engineer; cofounder of Apple Inc.
Andreessen, M.	Venture capitalist and software engineer; founder of Mosaic and Netscape; author of the Techno-Optimist Manifesto.
Huang, J.	Engineer and entrepreneur; CEO of NVIDIA Corporation.
Bengio, Y.	Computer scientist; renowned for his work on artificial neural networks and deep learning; key figure at the Center for AI Safety.
Bankman-Fried, S.	Founder of the FTX cryptocurrency exchange (collapsed in 2022); advocate for cryptocurrency technology and leading figure in the effective altruism movement. Convicted of fraud in 2023 and sentenced to 25 years in prison.
Hinton, G.	Computer scientist and cognitive psychologist; pioneering work on artificial neural networks; Nobel Prize winner in physics (2024).
Bostrom, N.	Philosopher; known for his work on existential risk, longtermism, and human enhancement; former director of the Future of Humanity Institute (Oxford), closed in 2023.
Harari, Y. N.	Historian and public intellectual; author of Homo Deus: A Brief History of Tomorrow (2017), which discusses technological future scenarios.
Horvitz, E.	Computer scientist; chief scientific officer at Microsoft Inc.; pioneer in AI and human-computer interaction.
Moskowitz, D.	Internet entrepreneur; cofounder of Facebook; primary funder of Open Philanthropy, a foundation that funds research and projects based on the principles of effective altruism.

The third empirical realm consists of newspaper reports in the time span between November 2022 and August 2024 by journalists covering chatbots or figures on the list commenting on their risks or achievements. Given the impressive amount of media coverage of the topic, I had to limit my scope and opted for the American newspaper The New York Times (NYT), an acclaimed daily newspaper that counts the third largest online visits (May 2024) among all news sites worldwide (Press Gazette, 2024). It holds most Pulitzer Prize winners and also features most subscribers, with 10.84 million (first half 2024; Miller, 2024). It can be considered a leading medium whose reports become disseminated broadly and internationally in other media broadcasting. I worked with the search prompts “CHATGPT,” “LLM,” and “CHATBOT,” leading to a total output of 97 articles for “CHATGPT,” 46 articles for “LLM,” and 723 articles for “CHATBOT.” I manually eliminated doubles and unrelated outputs like newsletters and focused on a final sample of articles that especially focused on chatbots’ abilities in the beginning of the release of ChatGPT and its aftermath, arriving at a final sample of 86 articles, including NYT interviews with the figures in Table 1.

The mode of analysis of the empirical material is hermeneutic, creating text codes that give consideration to the underlying rhetorical build-up of the hype. This comprises distinct metaphors, motifs, narratives, or cultural references, which are also present in the visual design of the company’s webpages where the chatbots are hosted. As hypes work with a play of emotions and impression management, I especially included motives that play with theatrical techniques (Goffman, 1959) and a “dramaturgical regime” (Oomen, Hoffman, & Hajer, 2022) of the chatbot hype, implying that its “performative imaginations are enacted” and can be deconstructed (Oomen et al., 2022, p. 259). In the forthcoming section, I dissect, going chronologically with the time of release of ChatGPT, the different hype building blocks that fabricated chatbots as a core institution to, as the narrative suggests, shape the future of humanity.

Analysis

Strategic Ignorance: Depicting LLM Chatbots as Knowledge Models

The beginning of the media frenzy about chatbots based on LLMs can be dated around winter 2022. LLMs are language-processing or -generating models that can process large amounts of text data (Bender, Gebru, McMillan-Major, & Shmitchell, 2021; Bommasani et al., 2021). It is not that LLMs had not existed before 2022,[4] but they simply had not received a lot of public media attention. Meta (2020) released the conversation bot BlenderBot in August 2022, after already having made the model and code public in April 2020, but BlenderBot gave wrong factual answers. It was shut down by the compliance department from Meta and became a flop for Zuckerberg’s company.

It was not until the then unknown company OpenAI released ChatGPT on November 30, 2022, that public and corporate interest sparked. ChatGPT actually had the same issues concerning hallucinations and biases like all the other LLMs to date, but OpenAI’s team simply released it anyway and used the public as a laboratory to flag malfunctioning and to train the model through user traffic. With the premature public launch, Open AI followed the first-mover strategy, willfully accepting potential severe risks for the public to secure market shares. The perception of ChatGPT in the releasing stage was crucial for the take-off of the chatbot hype. From an attention logic, to get a hype going, it is always easier to frame an initial dominant perception and to set an agenda than ex-post struggling and countering it, as rhetorically criticizing a phenomenon necessarily involves mobilizing the phenomenon.

The design and the properties of the interface to access and interact with ChatGPT favor a phenomenological experience that sparks curiosity and interaction. The play-and-answer game in a chat “what if” design (Heuser & Vulpius, 2024) easily anthropomorphizes the chatbot as a playful other, evoking self-reflection and empathy, known as the psychological ELIZA effect.[5] The initial user experience of ChatGPT, facing an opaque, but powerful, counterpart that gives plausible-sounding and creative answers, made it hard for many users to manage their expectations about its performance. Initial user experiences maneuver somewhere in between insecurity, excitement, and fascination—which, in turn, creates traction. ChatGPT appears to be some kind of human intelligence because it seems to have a hermeneutical understanding of context, giving witty and funny-sounding answers. It appears to have access to facts and knowledge, as if it “studied” or “contained” indexed knowledge like the entirety of Wikipedia, news articles, books, or scientific papers.[6] OpenAI also promoted this perception of an omniscient counterpart by luring the user on the landing page with the offer: “Ask me anything!” (Figure 1) (retrieved with a Wayback Machine), listing all kinds of domains that ChatGPT could be a helpful companion for, but ChatGPT never contained or accessed any information in its beginning phase.[7] It was trained with a large amount of text scraped from the Internet in 2021 (OpenAI, 2022). LLMs operate purely stochastically, mapping the statistical relationships between tokens (small parts of text) based on parameters (rules to process the tokens) to generate word sequences. Every word sequence is derived from predictions of the training data, making all text output always fictional (Bender et al., 2021; Hutson, 2024). Hence, LLMs are not designed to represent the world. There is no understanding by the artificial agent (chatbot) of the meaning of the output it creates. It is we humans who create that meaning. Given their functional logic, it is actually not surprising that LLMs have problems with the truth, such as historic facts, or the field of logic. They ‘hallucinate’,⁷ as they are not designed for these application domains. However, providers of these systems and known figures in the tech scene failed in making clear both the functionality and limitations of chatbots based on LLMs. To the contrary, they spurred the hype around seemingly omniscient and omnipotent chatbots.

Figure 1. ChatGPT mobile and web interface: “Ask me anything” landing page (ChatGPT, 2023).

Figure 2. Meta AI mobile interface: “Ask Meta AI anything” landing page[8] (Meta AI, 2023).

Figure 2. Meta AI mobile interface: “Ask Meta AI anything” landing page [8] (Meta AI, 2023).

The Weird and Eerie: Panicking About the Uncanny Side of Chatbots

The praise and media outcry ChatGPT received on its release were enormous. Influential figures in the tech industry overdid each other with praise in the first months. Elon Musk (2022) tweeted on X (194 million followers back then), “ChatGPT is scary good. We are not far from dangerously strong AI.” Former Microsoft CEO Bill Gates went so far as to declare a new epoch in history on his blog: “A new era, The Age of AI has begun. Artificial intelligence is as revolutionary as mobile phones and the Internet” (Gates, 2023, para. 1). Nick Bostrom even started to fabulate about AI chatbots having started the development toward general sentience, arguing in a NYT interview, “What if A.I. Sentience Is a Question of Degree?” (Bostrom & Jackson, 2023), and the popular historian and futurist Yuval Noah Harari warned in the print NYT edition with the headline “If We Don’t Master A.I., It Will Master Us” (Harari, Harris, & Raskin, 2023). Simultaneously, discussions in the media sparked speculation that high school language teachers would become obsolete because of chatbot performance (Herman, 2022), and student cheating would become impossible to detect (Nolan, 2023), urging universities to restructure exams.

This praise of ChatGPT fluctuated between hysteric emotions of panic and mesmerizing awe, entertaining the AI dualism of both redemption (LLMs going to solve all tedious problems) and doom (LLMs going to eradicate all jobs). A peak of this psychological insecurity, creating the attention frenzy, was reached when users and journalists discovered the dark side of LLMs: their seductive, manipulative, and “rogue” behavior, which was interpreted as dangerous and eerie. The NYT coverage substantially contributed to this perception when journalist Kevin Roose (2023) published the chat script with the Microsoft chatbot Bing, showcasing how the bot tried to convince him that he was unhappy in his marriage and should leave his wife to be with Bing instead. This was accompanied by other user testimonials reporting how LLMs would go rogue and refuse to be shut down (Kare 11, 2023). Below (Figure 3) are some NYT headlines from the start of the chatbot hype:

Figure 3. NYT article headlines addressing ChatGPT (Metz, 2022; Roose, 2023b, 2023c, 2023d).

In general, the perceived potency and humanlike agency of LLMs were both praised positively as a momentous revolution and technological breakthrough and negatively with the discovery of the uncanny side of LLMs. Both stimulated the media discourse about chatbots with emotional reactions, constantly creating new traffic feeding the hype. These speculations of an omnipotent agent or a manipulative, uncanny intelligence not only flooded a fictional space with speculations about what LLMs would possibly be able to do in the near future but also provoked a performative realm of necessity for action. Both corporate competition to OpenAI (in terms of market share) and politics (in terms of regulation) were on the spot to catch up with the perceived threat, which the unleashing of chatbots embodied.

The Battle: Staging a Spectacle of Competition Among Tech Giants

When OpenAI released ChatGPT, the competition was awakened to catch up with rival products. In 2023, Microsoft, which financed OpenAI as the biggest donor, incorporated ChatGPT in its search engine, Bing, and worked on releasing Copilot, an assistance bot for the office applications on its platform, Microsoft. Alphabet worked on Bard, which later became Gemini, and IBM on WatsonX. The media portrayed this rivalry among “BigTech giants” as a competitive race and entertained metaphors and scenarios of kick-off, catch-up, or survival.

This rhetoric points to a realm of action with leapfrogging the competition (studied in AI races in the geopolitical realm among nations; see Bareis & Katzenbach, 2021), raising the company’s potential, and mobilizing all capacities to rush for a promised future. The NYT wrote an indicative background article headlined in sensationalist wording: “Inside the A.I. Arms Race That Changed Silicon Valley Forever” (Weise, Metz, Grant, & Isaac, 2023). Here, even military metaphors are used to circumscribe a battle of tech giants. In the article, the established computer scientist and 2024 Nobel Prize winner of physics, and back-then Google employee, Geoffrey Hinton warned,

If you think of Google as a company whose aim is to make profits, [. . .] they can’t just let Bing take over from Google search. They’ve got to compete with that. When Microsoft decided to release a chatbot as the interface for Bing, that was the end of the holiday period. (Weise et al., 2023, para. 72)

Other articles by the NYT kept on entertaining the battle metaphor (Figure 4):

Figure 4. NYT article headlines addressing ChatGPT (Roose, 2023a; Roose & Pichai, 2023).

The battle metaphor was actually more harnessed by NYT framing than by the CEOs of the companies. For example, in an NYT interview, Alphabet CEO Pichai argues,

Sometimes I get concerned when people use the word “race” and “being first.” [. . .] We are definitely working with technology which is going to be incredibly beneficial, but clearly has the potential to cause harm in a deep way. And so I think it’s very important that we are all responsible in how we approach it. (Roose & Pichai, 2023, para. 16)

The news media depiction of the big tech competition is very much portrayed as an exciting spectacle feeding the hype. Readers and followers are entertained as if sitting in an arena to watch a fierce race to the top. The presented future trajectory enabled through this innovation overbidding opens a seemingly endless realm of technological possibilities reachable through competition. The spectacle invites an audience to follow a path to visit the never-before-visited or seemingly unachievable. This notion may be best captured by OpenAI’s CEO, Sam Altman, stating that AI’s benefits for humankind could be “so unbelievably good that it’s hard for me to even imagine” (Loizos & Altman, 2023, 21:20). Spectacles like the portrayed battle among big tech giants leave an audience in a strangely passive awe to observe great wonders they cannot control, but also cannot help watch, catering to psychological effects of both attraction and distance. However, being distant as actors, the spectacle could not function without the emotional part taking of an audience, rendering the audience very much complicit and necessary for the hype phenomenon to take place. Exactly by not questioning and not acting, but emotionally partaking in the image—or in the representation of a better or doomed future enabled through chatbots—a community of followership is crafted and united behind a hype. This phenomenon of both paralysis and complicity shows great similarity to what social critic Guy Debord (1967) described in Society of the Spectacle: “One cannot abstractly contrast the spectacle to actual social activity [. . .]. Lived reality is materially invaded by the contemplation of the spectacle [. . .] passive identification with the spectacle supplants genuine activity” (para. 8). Further, he wrote on the forging of community through gazing toward and contemplating on a future vision, “All that was once directly lived has become mere representation. [. . .] The spectacle is not a collection of images; rather, it is a social relation among people, mediated by images” (Debord, 1967, para. 4).

Crossing the Line of the Normal: Praising the Dualism of an Apocalypse and a Tech-Religious Calling

The frenzy around chatbots reached its peak when an open letter, signed by popular and established figures from Silicon Valley and AI experts, was published calling “all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4” (Future of Life Institute, 2023, para. 1). The letter, hosted on the webpage of the Future of Life Institute, rhetorically asks, “Should we develop nonhuman minds that might eventually outnumber, outsmart, obsolete and replace us? Should we risk loss of control of our civilization?” (Future of Life Institute, 2023, para. 2). This call for a moratorium on the development of AI was followed by an open letter from the Center for AI Safety (CAIS), where again public figures and AI experts argued that “mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war” (CAIS, 2023, para 2).

These concerns from the tech world also reached politics, as computer scientist Yoshua Bengio, signer of both of the aforementioned statements, was allowed to present those worries to the U.S. Senate twice, once in July 2023 (Bengio, 2023a) and once in December of the same year (Bengio, 2023b). EU Commission President von der Leyen also quoted the exact CAIS phrase about the “risk of extinction from AI” in her Speech of the European Union to the EU parliament in September 2023 (von der Leyen, 2023, para. 9).

The perceived overheated development of unleashed chatbots led to a doomsday apocalypse rhetoric in these statements. AI became depicted as a megalomaniac superpower, spinning scenarios of a power-seeking intelligence that could gain control over humanity’s critical sectors. Notions of a sudden hostile takeover, civilization chaos, and end-of-the-world armageddon sounded very much like science-fiction blockbusters—and the effect was maximum attention for chatbots unleashing this speculation. Emotionally, it caused worries and panic of not being able to control a runaway AI development potentially raging. A disruptive and accelerationist technological development is stylized as an unpredictable moment where a superintelligence bursts on humanity and rhetorically overthrows it like a revolution. However, instead of inviting a substantiated regulatory debate about the current risks of AI, the open letters set a different discourse. The apocalypse narrative silenced more political and structural regulatory questions about infrastructure control, AI models, and data ownership in the hands of a few private big tech players. Issues about the current democratic risks of misinformation or polarization through already synthetically produced content through LLMs on the Internet were also sidelined. LLMs caused a hype by entertaining an apocalypse and thereby exploited all attention resources for a highly speculative, catastrophic future. The very CEOs and tech-progress apologists who are responsible for developing the respective AI systems, using society as an open laboratory for their tryouts, became the ones stressing the catastrophic risks of AI. Psychologically, the narrative was very effective. Notwithstanding maximum attention, in the end, none of the companies paused their AI development or put down their chatbots—Italy simply banned ChatGPT because of privacy concerns for a month, just to make it accessible again with privacy-related changes.

Concluding Discussion: Interdependent Social Domains in the Hype Dynamic

The chatbot frenzy powerfully informs us how tech hype is triggered and nourished and how it gives prominence to a few powerful actors, who are given the authority to speak and to envision our societal future. This study reconstructed how chatbots based on LLMs got hyped into being. A detailed summary of the dramaturgy and the motifs distilled from the analysis can be found in the infographic (Figure 5).

Figure 5. The analysis motifs and their effects summarized.

There are many accomplices who take part in the hype production of chatbots: start-ups, such as OpenAI, that test their new innovations recklessly by putting them on the market without taking accountability for their effects on democracy, justifying this move as “gain[ing] experience with operating them in the real world” and “the best way to carefully steward AGI into existence,” as Sam Altman (2023, p. 11) puts it. It should be made transparent that many figures in Table 1 are followers of the ideology of tech accelerationism and proponents of the ethics of effective altruism. This creed of thinking postulates a positivist understanding of technology development and follows a strictly utilitarian and reductionist moral weighing between humans to “save” the future of the human race. In saving humanity, even extraterrestrial colonization is seen as necessary. Wenar (2024) tries to encapsulate this ethic with the following depiction:

Unborn generations could be worth a lot more than we are today, given population growth. What’s the point in deworming a few hundred kids in Tanzania when you could pour that money into astronautical research instead and help millions of unborn souls to escape Earth and live joyfully among the stars? (para. 86)

This speculative “as-if” (Nordmann, 2007) ethics escapes the scientific realm to be disproven, as its catastrophic thinking only operates in the highly speculative. Still, it is highly performative and impactful, calling for the need to establish heuristics of hype assessment and hype literacy (Belsunces & Bareis, 2025). Many tech billionaires care little about established democratic institutions and propagate their own ideology of tech-libertarian society, framing “regulation, democratic accountability, and rights-based governance as obstacles to be bypassed by their tech-ruled vision of the future” (Belsunces & Bareis, 2025, para. 14). This power accumulation of big tech itself shows strong analogies to an accelerating out-of-control agent. As Leggett (2021) turns the table,

Superintelligence? In short, we have created a corporate market machine that is now capable of manipulating and controlling individual humans, and that is infinitely better, already, at this task than any human is, or could hope to be. And we have given this machine the single, overarching goal of obtaining a return to capital. (p. 736)

This also puts into the picture among the regulators, who are not innocent in the hype production. As Stilgoe (2020) observes with hype,

Governments have also abdicated their responsibility [. . .]. Hype is a way of deliberately concealing the politics of technology and privatising progress. If we overinvest our hopes in new technologies, we underinvest in other necessary but less glamorous areas, including education, public health, infrastructure and maintenance. (p. 51)

Even worse, if governments envision tech futures themselves, they reify the same corporate tech imaginaries around AI instead of leading the way to an inclusive account where technology serves the public good—not to mention dubious political alliances (e.g., between Donald Trump and Elon Musk or Peter Thiel and J. D. Vance).

The hyped narrative of catastrophic AI tied to chatbots already shows a power shift, lock-in effects, and policy materializations. Just in October 2024, Yoshua Bengio was named chair to draft a code of practice on general purpose AI for the European Commission (Wold, 2024). In May 2025, Commission president von der Leyen stated at the annual EU budget conference, “We thought AI would only approach human reasoning around 2050. Now we expect this to happen already next year” (von der Leyen in Henning, 2025, para. 3), for which she was criticized by NGOs and scientists for hyping up AI. In the United States, the influence of doomsday apologists is even more striking. The aforementioned CAIS codrafted the Californian bill SB1047, which would have set up an oversight board, imposing safety model testing and legal liability for big tech companies. However, it was oddly framed, with existential risks, holding companies liable to “mitigate the risk of catastrophic harms from AI models so advanced that they are not yet known to exist” (Bauer-Kahan, 2024, p. 1). As tech journalist Merchant (2024) commented on the bill, “None of these [liabilities] are necessarily bad things, but for those of us who aren’t all that worried that the real threat of AI is that it will build a killer chemical weapon, its priorities seemed skewed” (p. 5). Big tech lobbied against the bill, which was seen as a potential blueprint for a national U.S. AI legislation. In addition, Democrats like Nancy Pelosi, former speaker of the U.S. House of Representative, voiced public opposition to a Democrat-sponsored bill—with success, as the bill was vetoed by Californian Democrat governor Newsom in August 2024 (Allyn, 2024).

As another building block in the game of hype production, there is the news media. Some of the NYT coverage of the chatbots catered to sensationalism and tried to entertain more than to inform the public critically. The NYT staged Valley CEOs with a prominent position to speak and often put up clickbaiting headlines, taking interview statements out of context as bold headlines. The newspaper rather stylized a spectacle of a battle among big tech players instead of reflecting on the consequences of their power positions. Most problematically, all of this spectacle consumes the limited attention spans of readers and hinders them from thinking about different futures and also about less Western, but maybe more global, ones like Afrofuturism or Sinofuturism—or, overall, about less technological and more social problem-solving trajectories. This points—last, but not least—the finger at citizens, the entertained audience as the complicit bystander, letting companies and politics get away with all of it. After all, a spectacle in an empty arena cannot be a spectacle.

Dissecting the actors in the production of the chatbot hype shows that hype depends on an interdependent dynamic to become a powerful societal phenomenon. For a hype to gain traction (i.e., being circulated widely across societal fora and stimulating investments), it needs an innovation trigger, as much as the hailing of charismatic leaders, media, loyal followers, the embracing of politics—and the complicity of a tacit audience. The findings are limited by not taking quantitative accounts of financial investments or confidence stock-marketing developments into consideration, another promising trajectory in tracing hypes. Here, quantitative accounts of controversy and network analysis could contribute to tracing the network power of single hypers and their outreach (cf. Marres et al., 2024).

A toolkit on how to counter hypes cannot be discussed here exhaustively, but if fallacious framing and highly speculative futures tied to a chatbot are the core issue, then it also opens trajectories for combating the problem (Belsunces & Bareis, 2025). Public advocates and policymakers must address the authority and credibility of knowledge production. This tactic follows another rationale than fighting a lost battle of fact checking, debiasing, and auditing rapidly increasing synthetical content, as proposed by current tech regulation such as the EU AI Act. An effective way to defeat hypers is to ignore them or, best, not give them the stage in the first place. Any engagement, no matter if positive praise or criticism on grounds of rationality or plausibility, is just feeding the attention machine. For hypers, the media credo “bad news is good news” also holds. If critique is to be launched, it must aim at the legitimacy of the speaking position.

Theoretically, this study shows that hype deserves more attention as an analytic concept in understanding the logics that drive our future at the crossroads of innovation, capitalism, media, and politics. Hypes are a truly modernist creed: They produce and need visions of tech progress; dynamics of opportunism and risk; and a simplified, innocent future liberated from the reflection of societal consequences to prosper.

References

Allyn, B. (2024, September 29). California Gov. Newsom vetoes AI safety bill that divided Silicon Valley. NPR. Retrieved from https://www.npr.org/2024/09/20/nx-s1-5119792/newsom-ai-bill-california-sb1047-tech

Altman, S. (2023, February 24). Planning for AGI and beyond. Open AI. Retrieved from https://openai.com/index/planning-for-agi-and-beyond/

Alvial-Palavicino, C. (2015). The future as practice. A framework to understand anticipation in science and technology. Tecnoscienza–Italian Journal of Science & Technology Studies, 6(2), 135–172.

Bareis, J., & Bächle, T. C. (2025). The realities of autonomous weapons: Hedging a hybrid space of fact and fiction. In T. C. Bächle & J. Bareis (Eds.), The realities of autonomous weapons (pp. 1–32). Bristol, UK: Bristol University Press.

Bareis, J., Roßmann, M., & Bordignon, F. (2023). Technology hypes: Practices, approaches and

assessments. TATuP-Zeitschrift für Technikfolgenabschätzung in Theorie und Praxis, 32(3), 11–16. https://doi.org/10.14512/tatup.32.3.10

Bareis, J., & Katzenbach, C. (2021). Talking AI into being: The narratives and imaginaries of national AI strategies and their performative politics. Science, Technology, & Human Values, 47(5), 855–881.

Barthes, R. (1992). Mythologies. New York, NY: Farram Straus and Giroux.

Bauer-Kahan, R. (2024, June 18). Safe and Secure Innovation for Frontier Artificial Intelligence Models Act. Assembly Committee on Privacy and Consumer Protection. Retrieved from https://apcp.assembly.ca.gov/system/files/2024-06/sb-1047-wiener-apcp-analysis_0.pdf

Beckert, J. (2016). Imagined futures: Fictional expectations and capitalist dynamics. Cambridge, MA: Harvard University Press.

Belsunces, A., & Bareis, J. (2025, August 21). Expanding hype literacy to protect democracy. Tech Policy Press. Retrieved from https://www.techpolicy.press/expanding-hype-literacy-to-protect-democracy/

Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big? . In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (pp. 610–623). https://doi.org/10.1145/3442188.3445922

Bengio, Y. (2023a, December 6). Written Testimony of Professor Yoshua Bengio. Presented before the U.S. Senate Forum on AI Insight Regarding Risk, Alignment, and Guarding Against Doomsday Scenarios. Retrieved from https://yoshuabengio.org/wp-content/uploads/2023/07/Written-Testimony-and-biography-of-Yoshua-Bengio_U.S.-Senate-Judiciary-Subcommittee-on-Privacy-Technology-and-the-Law_25_07_2023.pdf

Bengio, Y. (2023b, July 25). Written Testimony of Professor Yoshua Bengio. Presented before the U.S. Senate Judiciary Subcommittee on Privacy, Technology, and the Law. Retrieved from https://yoshuabengio.org/wp-content/uploads/2023/07/Written-Testimony-and-biography-of-Yoshua-Bengio_U.S.-Senate-Judiciary-Subcommittee-on-Privacy-Technology-and-the-Law_25_07_2023.pdf

Binder, W. (2022). Technology as (dis-)enchantment. AlphaGo and the meaning-making of artificial intelligence. Cultural Sociology, 18(1), 24–47. https://doi.org/10.1177/17499755221138720

Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, S., . . . Liang, P. (2021). On the opportunities and risks of foundation models. arXiv Preprint. arXiv:2108.07258

Bostrom, N., & Jackson, L. (2023) What if A.I. sentience is a question of degree?. New York Times. Retrieved from https://www.nytimes.com/2023/04/12/world/artificial-intelligence-nick-bostrom.html

Bordignon, F., Ermakova, L., & Noel, M. (2021). Over-promotion and caution in abstracts of preprints during the COVID-19 crisis. Learned Publishing, 34(4), 622–636. https://doi.org/10.1002/leap.1411

Borup, M., Brown, N., Konrad, K., & Van Lente, H. (2006). The sociology of expectations in science and technology. Technology Analysis & Strategic Management, 18(3–4), 285–298. https://doi.org/10.1080/09537320600777002

Bourne, C. (2024). AI hype, promotional culture, and affective capitalism. AI and Ethics, 4(3), 757–769. https://doi.org/10.1007/s43681-024-00483-w

Brock, D. C. (2019). Our censors, ourselves: Commercial content moderation. Los Angeles Review of Books, 25(07). Retrieved from https://lareviewofbooks.org/article/our-censors-ourselves-commercial-content-moderation/

Campolo, A., & Crawford, K. (2020). Enchanted determinism: Power without responsibility in artificial intelligence. Engaging Science, Technology, and Society, 6(0), 1–19. https://doi.org/10.17351/ests2020.277

Castro, A., & Belsunces, A. (2025). Cryptofinancial imaginaries: How neoliberal theories are materialized in the technical principles of cryptocurrencies. Journal of Cultural Economy, 18(5), 699–723. https://doi.org/10.1080/17530350.2024.2436869

Cave, S., & Dihal, K. (2019). Hopes and fears for intelligent machines in fiction and reality. Nature Machine Intelligence, 1(2), 74–78. https://doi.org/10.1038/s42256-019-0020-9

Cave, S., & ÓhÉigeartaigh, S. S. (2018). An AI race for strategic advantage: Rhetoric and risks. In Proceedings of the 2018 AAAI/ACM Conference on AI, Ethics, and Society (pp. 36–40). https://doi.org/10.1145/3278721.3278780

Center of AI Safety. (2023, May 30). Statement of AI risk. CAIS. Retrieved from https://www.safe.ai/work/statement-on-ai-risk

ChatGPT. (2023). Landing page. Retrieved from https://web.archive.org

Davis, A. (2013). Promotional cultures: The rise and spread of advertising, public relations, marketing and branding. London, UK: Polity.

Davis, E., & Aaronson, S. (2023). Testing GPT-4 with Wolfram alpha and code interpreter plug-ins on math and science problems. Arxiv Preprint. arXiv:2308.05713v2

Debord, G. (1967). Society of the spectacle. Marxists. Retrieved from https://www.marxists.org/reference/archive/debord/society.htm

Duff, A. S. (2016). Rating the revolution: Silicon Valley in normative perspective. Information, Communication & Society, 19(11), 1605–1621. https://doi.org/10.1080/1369118X.2016.1142594

Frankfurt, H. G. (2005). On bullshit. Princeton, NJ: Princeton University Press.

Future of Life Institute. (2023, March 22). Pause giant AI experiments: An open letter. Retrieved from https://futureoflife.org/open-letter/pause-giant-ai-experiments/

Galanos, V. (2023). To have done with the metaphor of summers and winters: Can AI and Internet history cure hype? Retrieved from https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4640305

Gates, B. (2023, March 21). The age of AI has begun. Retrieved from https://www.gatesnotes.com/The-Age-of-AI-Has-Begun

Geiger, S., & Gross, N. (2017). Does hype create irreversibilities? Affective circulation and market investments in digital health. Marketing Theory, 17(4), 435–454. https://doi.org/10.1177/1470593117692024

Goffman, E. (1959). The presentation of self in everyday life. London, UK: Penguin.

Grin, J., & Grunwald, A. (Eds.). (2000). Vision assessment: Shaping technology in 21st century society: Towards a repertoire for technology assessment. Berlin, Germany: Springer.

Harari, Y., Harris, T., & Raskin, A. (2023, March 24). Opinion: You can have the blue pill or the red pill, and we’re out of blue pills. Retrieved from https://www.nytimes.com/2023/03/24/opinion/yuval-harari-ai-chatgpt.html

Hepp, A. (2024). Curators of digital futures: The life cycle of pioneer communities. New Media & Society, 27(9), 5390–5409. https://doi.org/10.1177/14614448241253766

Hepp, A., Loosen, W., Dreyer, S., Jarke, J., Kannengießer, S., Katzenbach, C., & Schulz, W. (2023). ChatGPT, LaMDA, and the hype around communicative AI: The automation of communication as a field of research in media and communication studies. Human-Machine Communication, 6(1), 41–64.

Herman, D. (2022, December 9). The end of high-school English. The Atlantic. Retrieved from https://www.theatlantic.com/technology/archive/2022/12/openai-chatgpt-writing-high-school-english-essay/672412/

Heuser, M., & Vulpius, J. (2024). “Grandma, tell that story about how to make napalm again”: Exploring early adopters’ collaborative domestication of generative AI. Convergence: The International Journal of Research Into New Media Technologies. Advance online publication. https://doi.org/10.1177/13548565241285742

Henning, M. (2025, November 10). Stop overhyping AI, scientists tell von der Leyen. Euractiv. Retrieved from https://www.euractiv.com/news/stop-overhyping-ai-scientists-tell-von-der-leyen/

Hiltzik, M. (2025, September 5). The air begins to leak out of the overinflated AI bubble. Los Angeles Times. Retrieved from https://www.latimes.com/business/story/2024-09-05/the-air-begins-to-leak-out-of-the-overhyped-ai-bubble

Hockenhull, M., & Cohn, M. L. (2021). Hot air and corporate sociotechnical imaginaries: Performing and translating digital futures in the Danish tech scene. New Media & Society, 23(2), 302–321. https://doi.org/10.1177/1461444820929319

Hohmann, M. H., Barnett, A. G., King, N., & Connell, S. D. (2025). The evolution of scientific writing: An analysis of 20 million abstracts over 70 years in health and medical science. Scientometrics, 130(7), 3349–3366. https://doi.org/10.1007/s11192-025-05353-8

Hutson, M. (2024). How does ChatGPT “think”? Psychology and neuroscience crack open AI large language models. Nature, 629(8014), 986–988. https://doi.org/10.1038/d41586-024-01314-y

Kare 11. (2023, February 15). Testing the limits of ChatGPT and discovering a dark side [Video file]. Retrieved from https://www.youtube.com/watch?v=RdAQnkDzGvc

Kari, M., Lehtonen, M., Litmanen, T., & Kojo, M. (2023). Technology hype, promises, and expectations: The discussion on small modular reactors in the Finnish newspaper “Helsingin Sanomat” in 2000–2022. Zeitschrift für Technikfolgenabschätzung in Theorie und Praxis, 32(3), 41–47. https://doi.org/10.14512/tatup.32.3.41

Kotliar, D. M. (2025). Can’t stop the hype: Scrutinizing AI’s realities. Information, Communication & Society. Advance online publication. https://doi.org/10.1080/1369118X.2025.2531165

Large Language Model. (2025, November 8). Wikipedia. Retrieved from https://de.wikipedia.org/wiki/Large_Language_Model

Leaver, T., & Srdarov, S. (2023). ChatGPT isn’t magic: The hype and hypocrisy of generative artificial intelligence (AI) rhetoric. M/C Journal, 26(5). https://doi.org/10.5204/mcj.3004

Leggett, D. (2021). Feeding the beast: Superintelligence, corporate capitalism and the end of humanity. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society (AIES ’21) (pp. 727–735). New York, NY: Association for Computing Machinery. https://doi.org/10.1145/3461702.3462581

Linden, A., & Fenn, J. (2003). Understanding Gartner’s hype cycles: Strategic analysis report. Retrieved from http://ask-force.org/web/Discourse/Linden-HypeCycle-2003.pdf

List of Chatbots. (2025, November 8). Wikipedia. Retrieved from https://en.wikipedia.org/wiki/List_of_chatbots

Loizos, C., & Altman, S. (2023, January 18). StrictlyVC in conversation with Sam Altman, part two (OpenAI) [Video file]. Retrieved from https://www.youtube.com/watch?v=ebjkD1Om4uw

Mager, A., & Katzenbach, C. (2021). Future imaginaries in the making and governing of digital technology: Multiple, contested, commodified. New Media & Society, 23(2), 223–236. https://doi.org/10.1177/1461444820929321

Marres, N., Castelle, M., Gobbo, B., Poletti, C., & Tripp, J. (2024). AI as super-controversy: Eliciting AI and society controversies with an extended expert community in the UK. Big Data & Society, 11(2). https://doi.org/10.1177/20539517241255103

Martino, J. P. (2003). A review of selected recent advances in technological forecasting. Technological Forecasting and Social Change, 70(8), 719–733. https://doi.org/10.1016/S0040-1625(02)00375-X

Merchant, B. (2024, September 30). In California, no AI bill is safe. Blood in the Machine. Retrieved from https://www.bloodinthemachine.com/p/in-california-no-ai-bill-is.safe

Meta. (2020, April 29). BlenderBot. A state-of-the-art open source chatbot. Retrieved from https://ai.meta.com/blog/state-of-the-art-open-source-chatbot/

Meta. (2024, July 30). Review of fact-checking label and Meta AI responses. Retrieved from https://about.fb.com/news/2024/07/review-of-fact-checking-label-and-meta-ai-responses

Meta AI. (2023). Landing page. Retrieved from https://web.archive.org

Metz, C. (2022, December 10). The new chatbots could change the world: Can you trust them? New York Times. Retrieved from https://www.nytimes.com/2022/12/10/technology/ai-chat-bot-chatgpt.html

Michalec, O. (2025). Models vs infrastructures? On the role of digital twins’ hype in anticipating the governance of the UK energy industry. Environmental Science & Policy, 168, 104041. https://doi.org/10.1016/j.envsci.2025.104041

Miller, H. (2024, August 7). New York Times beats estimates as digital subscriptions rise. Bloomberg. Retrieved from https://www.bloomberg.com/news/articles/2024-08-07/new-york-times-beats-estimates-adds-300-000-digital-subscribers

Milne, G. (2020). Smoke & mirrors: How hype obscures the future and how to see past it. Abingdon, UK: Hachette UK.

Musk, E. (2022, December 3). ChatGPT is scary good. We are not far from dangerously strong AI [Tweet]. Retrieved from https://x.com/elonmusk/status/1599128577068650498

Natale, S., & Ballatore, A. (2020). Imagining the thinking machine: Technological myths and the rise of artificial intelligence. Convergence: The International Journal of Research Into New Media Technologies, 26(1), 3–18. https://doi.org/10.1177/1354856517715164

Nolan, B. (2023, January 14). Two professors who say they caught students cheating on essays with ChatGPT explain why AI plagiarism can be hard to prove. Business Insider. Retrieved from https://www.businessinsider.com/chatgpt-essays-college-cheating-professors-caught-students-ai-plagiarism-2023-1?r=US&IR=T

Nordmann, A. (2007). If and then: A critique of speculative NanoEthics. NanoEthics, 1(1), 31–46. https://doi.org/10.1007/s11569-007-0007-6

Oomen, J., Hoffman, J., & Hajer, M. A. (2022). Techniques of futuring: On how imagined futures become socially performative. European Journal of Social Theory, 25(2), 252–270. https://doi.org/10.1177/1368431020988826

OpenAI. (2022, November, 30). Introducing ChatGPT. Retrieved from https://openai.com/index/chatgpt/

Palavicino, C. A., & Konrad, K. (2015). How technology consultants assess the graphene and 3D printing hype. Retrieved from https://www.academia.edu/download/69160155/download.pdf

Pocock, K. (2024, April 3). ChatGPT does have access to the internet, but not for all users. PC Guide. Retrieved from https://www.pcguide.com/apps/chatgpt-internet-access/

Powers, D. (2012). Notes on hype. International Journal of Communication, 6, 857–873. Retrieved from https://ijoc.org/index.php/ijoc/article/view/1441

Press Gazette. (2024, June 25). Top 50 biggest news websites in the world in November: New York Times is fastest-growing year-on-year but drops out of top three. Retrieved from https://pressgazette.co.uk/media-audience-and-business-data/media_metrics/most-popular-websites-news-world-monthly-2/

Ramiller, N. C. (2006). Hype! Toward a theory of exaggeration in information technology innovation. Academy of Management Proceedings, 2006(1), A1–A6. https://doi.org/10.5465/ambpp.2006.27169062

Rip, A. (2006). Folk theories of nanotechnologists. Science as Culture, 15(4), 349–365. https://doi.org/10.1080/09505430601022676

Roose, K. (2023a, February 3). How ChatGPT kicked off an A.I. arms race. New York Times. Retrieved from https://www.nytimes.com/2023/02/03/technology/chatgpt-openai-artificial-intelligence.html

Roose, K. (2023b, February 16). Bing’s A.I. chat: “I want to be alive. 😈.” New York Times. Retrieved from https://www.nytimes.com/2023/02/16/technology/bing-chatbot-transcript.html

Roose, K. (2023c, February 16). A conversation with Bing’s chatbot left me deeply unsettled. The New York Times. Retrieved from https://www.nytimes.com/2023/02/16/technology/bing-chatbot-microsoft-chatgpt.html

Roose, K. (2023d, February 17). The online search wars got scary. Fast. New York Times. Retrieved from https://www.nytimes.com/2023/02/17/podcasts/the-daily/the-online-search-wars-got-scary-fast.html

Roose, K., & Pichai, S. (2023, March 31). Google C.E.O. Sundar Pichai on the A.I. moment: “You will see us be bold.” The New York Times. Retrieved from https://www.nytimes.com/2023/03/31/technology/google-pichai-ai.html

Seidensticker, B. (2006). Futurehype: The myths of technology change. Berkeley, CA: Berrett-Koehler Publishers.

Sharma, A., & Grant, D. (2011). Narrative, drama and charismatic leadership: The case of Apple’s Steve Jobs. Leadership, 7(1), 3–26. https://doi.org/10.1177/1742715010386777

Stilgoe, J. (2020). Who’s driving innovation? New technologies and the collaborative state. Cham, Switzerland: Palgrave Macmillan.

Strange, M. (2024). Three different types of AI hype in healthcare. AI and Ethics. Advance online publication. https://doi.org/10.1007/s43681-024-00465-y

Thompson, J. (2010). Merchants of culture. London, UK: Polity.

Thubron, B. (2023, February 3). ChatGPT adds 100 million users in two months, making it the fastest-growing “app” ever. TechSpot. Retrieved from https://www.techspot.com/news/97486-chatgpt-adds-100-million-users-two-months-making.html

Vasterman, P. (2005). Media-hype: Self-reinforcing new waves, journalistic standards and the construction of social problems. European Journal of Communication, 20(4), 508–530. https://doi.org/10.1177/0267323105058254

von der Leyen, U. (2023, September, 13). State of the Union 2023 Address by the European Union Commission President. Retrieved from https://ec.europa.eu/commission/presscorner/detail/ov/speech_23_4426

Wadhwani, R. D., & Lubinski, C. (2025). Hype: Marker and maker of entrepreneurial culture. Journal of Business Venturing, 40(2), 106455. https://doi.org/10.1016/j.jbusvent.2024.106455

Weise, K., Metz, C., Grant, N., & Isaac, M. (2023, December 5). Inside the A.I. arms race that changed Silicon Valley forever. Retrieved from https://www.nytimes.com/2023/12/05/technology/ai-chatgpt-google-meta.html

Wenar, L. (2024, March 27). The deaths of effective altruism. Wired. Retrieved from https://www.wired.com/story/deaths-of-effective-altruism/

Wernick, A. (1991). Promotional culture: Advertising, ideology, and symbolic expression. London, UK: SAGE Publications.

Wien, C., & Elmelund-Præstekæker, C. (2009). An anatomy of media hypes: Developing a model for the dynamics and structure of intense media coverage of single issues. European Journal of Communication, 24(2), 183–201. https://doi.org/10.1177/0267323108101831

Wold, J. W. (2024, October 9). Academics to chair drafting the Code of Practice for general-purpose AI. Retrieved from https://www.euractiv.com/section/tech/news/academics-to-chair-drafting-the-code-of-practice-for-general-purpose-ai/

Woznica, M. (2022). Stage performances as means for linking sociotechnical imaginaries

and projective genres in the discourse around urban air mobility. European Journal of Futures Research, 10(1), 12. https://doi.org/10.1186/s40309-022-00198-3

[1] Gartner Consulting created the famous depiction of the hype circle (Linden & Fenn, 2003), which, albeit its visual presence, has been criticized as “folk theory,” lacking any empirical basis (Rip, 2006).

[2] See the 2024 founded research collective around hype studies and its activities: https://hypestudies.org/.

[3] Here, I clearly take a different route than Vasterman (2005), who argues that “‘exaggeration’ and ‘distortion,’ are not suitable for an analytic definition of media-hype” (p. 512). To the contrary, I think they lie at the core of the phenomenon.

[4] See here the frequently updated Wikipedia list for an overview of all existing chatbots and LLMs (“Large Language Model,” 2025; “List of Chatbots,” 2025).

[5] In 1966, the computer scientist Joseph Weizenbaum fed his chatbot ELIZA with the DOCTOR script, imitating a Rogerian psychotherapist. ELIZA was a very rudimentary chatbot, programmed to simply rephrase patients’ answers as backfeed questions. Weizenbaum was struck when he observed that his chatbot elicited very emotional and intimate responses from his probands.

[6] However, attempts to connect ChatGPT to search engines like Wolfram Alpha or access databases have so far not been successful in eliminating errors and hallucinations (Davis & Aaronson, 2023). Still in July 2024, Meta had to apologize for its chatbot META AI falsely stating that the assassination attempt on Donald Trump had never happened (Meta, 2024).

[7] GPT-3 can retrieve data from the Internet with API plugins since May 2023 (Pocock, 2024). Since the beginning of 2025, many chatbots offer quick-time access to web content, retrieving information from sources like news webpages.

[8] The landing pages of Le Chat, WatsonX, and Gemini are designed with a more sober aesthetic, avoiding anthropomorphic elements and exaggerated promises.