Short abstract:

Hyped expectations in the digital economy are no longer the same as once, especially as depicted in ‘first-wave’ studies surrounding the Internet boom and bust. Industry analysts undergird the operation of the digital economy and its reliance on hyped expectations.

Long abstract:

Our title, ‘After Hype’, does not suggest that hype has gone away. Instead, we argue that hyped expectations are no longer the same as once, especially as depicted in ‘first-wave’ studies surrounding the Internet boom and bust. Initially, the digital economy's various evaluative frameworks were rudimentary and inchoate. Comparing the development of digital ventures today with the development of Internet dotcoms from just a couple of decades ago (Garud et al., 2019), we have been struck by the number of specialised intermediaries that now steer hype in some regard. Indeed, one of the lesser-discussed issues following the bursting of the Internet bubble has been the emergence and rapid growth of a small but powerful class of actors – the ‘industry analyst’ – who have, in some respects, ‘colonised’ discussions of hype in the digital economy in that they are widely seen as the ‘appropriate’ (Suddaby & Greenwood, 2001) knowledge provider for this phenomenon. These firms have expertise in finding hype, describing it, visualising it, and even amplifying and circulating it. They have also created a (very large) audience for their services/knowledge. In response to the tendency for promising innovations to be surrounded by hyped expectations, they began to subject venture narratives to detailed scrutiny and evaluation as a matter of routine. Our study shows that, as a result, industry analysts have come to play a pivotal role in how hyped expectations are created and circulated. Their work undergirds the operation of the digital economy and its reliance on hyped expectations.

Short abstract:

This paper offers a computational case study of the relationships between how heath AI is measured and promotional language use in biomedical reporting. The results offer insights into discovery-justification interactions and can inform prospective approaches to addressing hype in health AI.

Long abstract:

Hyperbole in biomedical reporting has driven an overly enthusiastic embrace of these AI technologies. For example, premature of adoption of the Epic sepsis model has been linked to diminished outcomes in hospitals that predominantly serve marginalized populations. While there has been much discussion of hyperbole as a framework for addressing these issues, dominant conceptions of hyperbole rely on problematic notions of correspondence epistemology and discovery/justification extricability. This leads to research designs that evaluate hype retrospectively, assessing various correspondences such as the fit between reporting enthusiasm and underlying AI performance or mismatches between causal language and conducted statistical tests. Given the manifest harms of AI hype, a prospective orientation is necessary. To that end, this study offers an extended computational case study of promotional language in health AI. A supervised machine learning model was used to identify promotional language in a random sample of 1200 health AI abstracts drawn from PubMed, and quasi-Poisson regression was used to assess the relationship between performance metric selection and promotional language frequency. The results indicate that the use of certain computer science metrics predicts higher rates of promotional language when compared to more traditional biomedical measures. The computational case method here offers aggregate-level insights into discovery-justification interactions, and in so doing provides a foundation for future research supporting prospective approaches to addressing hype in health AI. Increased knowledge of which methods and metrics are more likely to lead to hype in health AI can provide an evidence-based foundation for early intervention.

Short abstract:

We empirically analyse how ‘AI’ is operationalised in German news and identify a strong disconnect between a future-oriented and promissory discourse and a nuanced contextual discourse. Based on this empirical starting point, we discuss how larger pro-innovation discourses shape the German AI hype.

Long abstract:

As a technical phenomenon that has generated over 7,500 articles in four years of German news reporting, the topic of ‘Artificial Intelligence’ certainly seems to qualify as ‘hype’. Yet, what the signifiers of both AI and hype exactly point to is a contentious and ongoing debate (Intemann 2022; Suchman 2023). In this paper, we qualitatively explore what kind of discursive object(s) ‘AI’ emerges as in German media reporting, thus offering an empirical starting point for conceptual debates around hype.

With our national focus, we contribute to a small body of research on AI in the German news, so far dominated by quantitative studies. Our approach allows us to assess how specifics of the German context, for instance the prevalence of engineering-based industries, influence, and shape AI discourses, thus studying hype as a context-specific phenomenon.

We conducted a discourse analysis examining how leading German newspapers discuss AI over four years, focusing on how articles construct AI and imbue it with meaning. Our analysis identifies two interrelated but distinct discourses around AI, situated on a continuum between future-oriented and promissory generalisation and nuanced, use-case specific particularity. We refer to them as discourses of AI-in-general and AI-in-particular (Michael 1992).

Our empirically grounded conceptual work speaks to the topic of this panel as it allows us to trace how ‘hype’ on AI emerges in specific contexts. It can thus serve as an invitation for joint reflections between STS scholars and AI practitioners on what it may mean to communicate responsibly in a hyped environment.

Short abstract:

Promising in science fosters cooperation but often leads to broken promises due to uncertainty and incentives to overpromise. Despite awareness, these practices persist, fueling hype and disappointment. I propose a query to track promises to understand why these issues are not corrected over time.

Long abstract:

One prominent activity contributing to hype is that of promising. Promising is a regular and often expected feature of the scientific endeavour, and with good reason – promising enables mutual confidence and makes cooperation possible. However, due to the inherent uncertainty of science, promises are often broken. In addition, scientists have much to gain in the short term by overpromising. These two factors combined lead to a high rate of broken promises in science.

Even though both scientists and funders seem aware of the prevalence of broken promises, current practices seem to continue unabated. The expectations following promises continue to contribute to the emergence of hype, with broken promises contributing to the almost inevitable disappointment following long periods of hype. I therefore propose the empirical, longitudinal study of promises, to better understand why problematic promising practices are not corrected over time. I present preliminary results of the development and application of a search query which allows users to search for predictions in a given corpus. For each prediction, the date when the prediction was made and other relevant metadata are given. The user can subsequently manually determine whether a prediction was a promise or not and, in combination with the metadata, analyse when a promise first arose and how it developed over time. This lays the foundation for further analysis of whether unrealistic promises are corrected over time and if not, which mechanisms are in place that incentivize exaggerated promises.

Short abstract:

This paper explores the recent resurgence of AI in corporate environments. As AI was stated as a strategic priority, many firms turned to consulting firms for implementation. We examine the role of these market intermediaries in spreading and operationalizing the AI hype over the past decade.

Long abstract:

This paper examines the resurgence of artificial intelligence (AI) since the mid-2010s, marked by its integration as a strategic priority in major corporations seeking to modernize their internal processes. Many of these companies sought the expertise of consulting firms to realize their AI ambitions, highlighting the pivotal role of these firms as 'market intermediaries' (Bessy, Chauvin; 2013). Our study focuses on how consulting firms have influenced the diffusion and practical operationalization of AI hype in enterprises over the past decade.

There has been a progressive evolution in corporate engagement with AI, moving from initial trial-and-error approaches that often seemed more trend-driven than strategically grounded, to a more mature and discerning approach. This shift is often encapsulated by the industry saying, "We don't do AI just for the sake of AI." This study critically analyzes the role of consulting firms in this evolving landscape. These firms face the paradox of capitalizing on the AI hype while advising their clients against blindly following trends.

We explore how consulting firms navigate this hype, often cautioning against the perils of trend-chasing while simultaneously opening up new market opportunities. We also discuss their role in fostering the emergence of an “AI governance” in companies to move beyond proof-of-concept stages. This paper argues that consulting firms play a crucial role in operationalizing the hype within organizations, providing meaning, tools, and a tangible form. We offer insights into how these actors shape and translate the AI hype into actionable strategies within the corporate sector.

Short abstract:

Focusing on the nanomaterial graphene and the University of Manchester’s ‘Graphene City’ vision, this paper will explore the role of architecture in sustaining, expanding and mobilizing ‘hype’ in order to produce obdurate networks of 2D material research and development.

Long abstract:

As with many scientific breakthroughs (Nowotny and Felt, 1997), the isolation of graphene at the University of Manchester in 2004 was met with a frenzy of activity, as scientists, engineers and entrepreneurs looked to mobilise and exploit the new discovery. Like many nanotechnologies (Hessenbruch 2004; Selin, 2007; McGrail, 2010), graphene has given rise to expectations and visions surrounding potential future technological applications, from engineering to aerospace. However, this paper will outline that, at the University of Manchester, the hype surrounding graphene also gave rise to an urban vision: Graphene City. Attempting to reimagine Manchester as a ‘Silicon Valley of graphene’, the vision would be centred around three new state of the art research facilities – the National Graphene Institute (NGI), the Graphene Engineering Innovation Centre (GEIC) and the Henry Royce Institute (HRI).

This paper will explore the role of architecture in sustaining and expanding the hype around graphene, and ultimately attempting to consolidate Manchester as a global node in graphene and 2D materials research – attempting to give rise to an obdurate network of graphene expertise, embedded within the city which would endure long after the ‘peak’ of graphene’s hype. Drawing on semi-structured interviews and an ethnographic study conducted at the GEIC in 2019, this paper will explore the ways that these new developments looked to reframe both the city of Manchester and its scientific communities as a ‘critical mass’ for the commercialisation of 2D material technologies, and establish itself as ‘the home of graphene’.

Short abstract:

Hype is back in public and scholarship conversation. To understand the understudied phenomenon, our paper suggests that hype tools, the tools that analyse hype, could provide valuable toolkits to study hype and set up promising agendas which future empirical inquiries might investigate.

Long abstract:

We propose three reasons why studying hype tools provides an invaluable way of approaching hype. First, we claim that hype tools provide a space for researchers to study hype, a phenomenon that is difficult to capture directly. Second, we point out that hype tools usually provide a context that helps anchor the meaning of hype and makes hype communicable. Finally, we argue that studying the mechanism that evaluates hype is studying hype itself. Our first contribution is providing a multidimensional understanding of the rationales operating underneath hype. Our second contribution links closely to our first contribution. We identify and articulate important sociological dimensions of hype tools. Through our exploration of the intricacies of hype tools and their role in capturing hype and channelling it to various stakeholders in the digital economy, our aim is to inspire future research that bridges diverse research traditions and disciplines, ultimately advancing our understanding of hype and hype tools. We propose that there are a number of promising fields that need future research attention. First, we suggest a shift from a result-based view to a practice-based view on hype tools. Second, we propose that the distinctive context of hype tools presents a fertile ground for researchers intrigued by the realm of multimodality. By encouraging interdisciplinary conversation, we hope to foster new insights and perspectives that contribute to the ongoing wave of hype management research.