Tech Tensions
Artificial Intelligence and Generative Media View Digital Media
Paper Presentation in a Themed Session Eric Freedman
This paper is focused on the role of artificial intelligence in generative media and lays a foundation to examine generative practices across a number of distinct media production fields, all of which have developed a common dependence on underlying AI frameworks. The concept of generative media, media that are created with and activated by a wide range of intelligence systems whose algorithms perform transformative and generative operations on collocated data, invokes the tensions between customization and control, automation and creative expression. Generative media have become popularized through both commercial and open access text to image generators such as Midjourney and Craiyon that use custom text prompts to generate synthetic images from a networked database. These tools are grounded in the research and development imperatives of OpenAI and its natural language systems as well as the more expansive goals of artificial general intelligence. These interactive communicative systems contour image creation to the semantics of their underlying algorithms and the familiar logic of a search engine; and as semi-autonomous, generative systems they call to question the relative freedom that exists in media that are fundamentally transactional. The goal of this paper is to understand the expanding industrial contours of creative programming and the impact of AI-driven image work on popular media culture. This paper does so through a critical analysis of the text to image generators Midjourney and Craiyon, their impact on visual culture, and their position within the grander industrial goals of OpenAI that foreground the economic values of autonomous systems.
A Biosemiotic Approach to the Problem of Trust in Artificial Intelligence Systems: Trust Is an Imputed Property of Symbols that Depends on the Pragmatic Consequences of Our Uses View Digital Media
Paper Presentation in a Themed Session Anderson Vinicius Romanini
From a logical point of view, trust is a general attribute that can only be predicated of complex symbols. It is in the accumulation of socially shared experiences within a community that uses artificial intelligence systems that trust appears as a general feeling resulting from the consequences produced by the social use of these systems. Peirce's semiotics and his original version of pragmatism can offer a relevant contribution to this debate, especially when we adopt the perspective of biosemiotics - the branch of semiotics that studies the phenomena of mind and life. For biosemiotics, living systems build internal representations about the external states of the environment where they live and make choices to act/interact based on the degree of confidence they have in 1) the pragmatic truth of their representations and 2) the predictability of the answers they collect from the environment. The pragmatic maxim, which establishes the meaning of a symbol from the consequences that it would be able to produce in the community that adopts it, places trust as an interactional and long-lasting property, which involves both past evidence and future expectations. Intelligent systems need to incorporate the temporal and teleological dimension that is expressed in conditional logical feelings, if not counterfactual.
Light in a Digital Black Hole: Exploration of Emergent Artificial Intelligence Journalism in Nigeria View Digital Media
Paper Presentation in a Themed Session Farooq Kperogi
Artificial intelligence journalism has been incorporated into the professional routines of the institutional, mainline news media in the West for a little over a decade, but it is only just now being slowly adopted in the rest of the world. This study deploys a combination of case-study research and semi-structured in-depth interviews with senior editors in Nigeria to explore the state of artificial intelligence journalism in Nigeria, Africa’s most populous country and one of its leading adopters of emerging technological innovations, and thereby inspire an expansion of the disciplinary conversations about the influence of robotic journalism in news gathering, production, and dissemination processes, especially in parts of the world that are symbolically peripheral and technologically marginal but nonetheless important parts of the global news ecology. It argues that the burgeoning embrace of robotic journalism in Nigeria and the professional anxieties this emergent reality is activating among journalists in the country have not been captured in the scholarly literature. It also discusses the implications of leapfrog innovation and the routinization of artificial intelligence reportorial practices in a digitally backward country.
Platform Research with Instagram Data : An Overview of Analytical Approaches, Digital Methods, and Research Ethics in Times of APICalypse
Paper Presentation in a Themed Session Patrick Nehls, Yannik Peters, Caja Carola Thimm
In recent years, the photo and video sharing platform Instagram has become one of the most used social media worldwide. Due to practical research barriers associated with the data collection, Instagram still seems to be underrepresented in (empirical) platform research. This paper therefore highlights the various ways in which Instagram data can be collected. First, we introduce the media grammar of Instagram. This approach helps to differentiate and theoretically frame the possibilities of collecting Instagram data presented below. In terms of data collection, a distinction is made between four different collection strategies: a) manual collection, b) free programming languages, c) tools of the Meta group and d) commercial tools. These are compared on the basis of relevant criteria (including media grammar, survey form, contribution levels, completeness, programming skills). It is shown that different data collection strategies may be appropriate depending on the research question and method. The aim of our study is to make a contribution to facilitating access to Instagram data and to support researchers in collecting Instagram data. Using our own data collection on the topic of COVID-19 on Instagram in Germany, we also present an example of an approach that combines several of these tools. Furthermore, the paper also takes a closer look at the ethical and legal perspectives of research.
