Artificial Intelligence (AI), while widely used, lacks a universally accepted definition. Recent advancements, particularly the release of ChatGPT, have significantly increased public interest and debate. ChatGPT's ability to perform various tasks, coupled with its limitations and potential risks (e.g., generating false information), has fueled public discourse, further intensified by calls for AI development regulation. This study investigates how ChatGPT and AI are represented in UK news media headlines from January to May 2023, utilizing Agenda-Setting theory (which emphasizes the media's role in shaping public opinion) and Framing theory (which focuses on how the media presents issues). The research aims to provide a rapid assessment of AI's portrayal in the media and analyze headline patterns over time.

Existing research highlights the use of newspaper corpora to study cultural and psychological phenomena, including technological change. Studies on newspaper headlines have explored their linguistic features, their role as attention-grabbing devices, and their potential for misrepresentation. Previous work has analyzed headlines to gauge sentiment (e.g., in COVID-19 reporting) and to understand political biases in AI coverage (Brennen, 2018). While research exists on AI's future role in journalism, studies specifically examining AI's representation in media, particularly regarding ChatGPT and LLMs, are limited.

The study collected 671 headlines from UK news media outlets (both print and online) using Lexis Library News Search, focusing on keywords like 'ChatGPT', 'AI', and 'Artificial Intelligence'. Headlines were included if the keywords were in the headline and the content was related to AI, ChatGPT, or LLMs. Articles with embedded links were excluded. Outlets with fewer than 20 articles were grouped into an 'Other' category, resulting in seven categories: The Times, The Independent, The Guardian, Other, The Daily Mail, The Daily Star, and The Telegraph. Political leanings of outlets were categorized based on the UK Government's National Readership Survey (YouGov, 2017). Inductive thematic analysis, a flexible method suited for analyzing headlines, was used to identify themes. Braun and Clarke's (2006) six-step approach was followed, starting with familiarization, identifying shared patterns, assigning codes, refining codes iteratively, and finally, developing overarching themes. The analysis also investigated the dispersion and frequency of articles across time and outlets.

Six main themes emerged from the headline analysis: 'Impending Danger' (37%, emphasizing potential societal damage from AI), 'Explanation/Informative' (26%, explaining AI basics and functions), 'Negative Capabilities of AI or ChatGPT' (11%, describing AI's failures or errors), 'Positive Capabilities of AI or ChatGPT' (11%, highlighting AI's benefits), 'Humorous/Comedic' (4%, using AI for entertainment), and 'Experimental Reporting' (4%, reporting on AI's experimental use). A small percentage (5%) of headlines were unclassified. Headline production increased significantly from January to May 2023, with a fivefold increase in May. 'Impending Danger' and 'Explanation/Informative' were the most prevalent themes across all outlets, regardless of political leaning. While some subtle differences between right and left-leaning outlets emerged, this was not as pronounced as found in previous research by Brennen (2018). Examples of headlines illustrating each theme were presented in tables.

The study's findings contribute to understanding how AI and ChatGPT are discursively represented in UK news media. The prominent 'Impending Danger' and 'Explanation/Informative' themes can be interpreted through Agenda-Setting and Framing theories. The frequent portrayal of AI as a potential threat may be influencing public perception and setting a societal agenda calling for AI regulation. The framing of AI narratives significantly shapes public understanding and attitudes. The 'Explanation/Informative' frame, while offering education, may inadvertently highlight the technology's complexity, creating uncertainty. The 'Impending Danger' frame likely aims to generate engagement through sensationalism. While some differences were observed in framing across political leanings, the overall representation of AI in the media is complex and multifaceted, beyond a simplistic dichotomy.

This study offers a valuable snapshot of AI representation in UK news media headlines, revealing a complex narrative oscillating between potential and peril. The increase in AI-related headlines underscores the growing public interest and the media's significant role in shaping perceptions. Future research should expand the timeframe, geographical scope, and delve into the interaction between media framing and public opinion. This would provide a more comprehensive understanding of the media's influence on public attitudes and inform strategies for fostering balanced and informed discussions about AI.

The study's limitations include the restricted timeframe (five months in 2023), potential biases in the inductive thematic analysis, and the focus on the UK media landscape, limiting generalizability to other contexts. Future research could broaden the temporal and geographical scopes to account for these limitations and provide a more global perspective.