Regulating AI? Exploring Czech Media Imaginaries with CADS

Description

Artificial Intelligence (AI) has recently raised enormous public and media concerns. This trend accelerated with the Chat-GPT release at the end of 2022. Among other topics where AI currently arouses public controversy, the issue of regulation is particularly prominent. As with any new technology, the future of AI depends on highly mediated public perception (Chuan et al. 2019). This case study focuses on AI regulation representation in the Czechia media and covers the period of discussions about the EU’s “AI Act” (3/2018-12/2023). It focuses on online mainstream media as one of the primary local sectors for expressing public opinion.
Czechia embodies the specific controversies of Eastern Europe: a significant mistrust of the EU aligning with a negative view of politics and legislation (Havlík and Hloušek 2017, p. 88-90); conversely, there is a predominantly positive outlook on digital technologies development (European Commission 2021). Using Corpus Approaches to Discourse Studies (CADS) (Baker et al. 2008) tools, a corpus of 675 articles (659 388 tokens) is explored. CADS allows the investigation of the aggregate effects of language, highlighting typical
discursive patterns. It combines corpus linguistics with traditional CDA while reflecting on its critiques (Orpin 2005). It mitigates mainly issues of data representativeness and interpretative transparency.
This study investigates media representations using the concept of “sociotechnical imaginaries” (SI), understood as “collectively held,
institutionally stabilised, and publicly performed visions of desirable futures” (Jasanoff and Kim 2015, p. 4). Current research on imaginaries is plagued by ambiguity (Richter et al. 2023). To escape this vicious circle, I apply Andrea Sau’s (2021, p. 10) three-level imaginary and operationalise it in two steps.
Firstly, I explore distinct imaginary layers focusing on collocations of the topic-related keywords (selected by frequency), contextualised within concordances. I reveal social commentary of the imaginary (1), which portrays AI as a speedily developing technology that needs to be regulated to avoid possible risks (dichotomy threat/opportunity). Despite the general scepticism, the vision of the future (2) is tied to the AI Act as proof of the EU’s ability to establish global regulation standards and express its digital identity. The general goal is to create a single digital market and challenge the digital superpowers (China, USA). Lastly, joint EU action, unity, and tech sector investments are the envisaged means to bring about this future (3).
Additionally, I focus on keywords of sub-corpus (426 994 tokens) covering the Chat-GPT period to reveal characteristics of the current imaginary. It confirms prevailing industry-led debate and domination of private actors (tech corporations and leaders) (Richter et al. 2023), even regarding concerns.

Period	19 Apr 2024
Event title	XVIII Fórum de Partilha Linguística | 18th Forum of Linguistic Sharing
Event type	Conference
Location	Lisboa, PortugalShow on map
Degree of Recognition	International