They Think AI Can Do More Than It Actually Can: Practices, Challenges, & Opportunities of AI-Supported Reporting In Local Journalism

TL;DR

The findings reveal that local journalists do not fully leverage AI's potential to support data-related work, and provide recommendations for improving AI-supported reporting in the context of local news, grounded in the journalists'socio-technical perspective and their imagined AI future capabilities.

Abstract

Declining newspaper revenues prompt local newsrooms to adopt automation to maintain efficiency and keep the community informed. However, current research provides a limited understanding of how local journalists work with digital data and which newsroom processes would benefit most from AI-supported (data) reporting. To bridge this gap, we conducted 21 semi-structured interviews with local journalists in Germany. Our study investigates how local journalists use data and AI (RQ1); the challenges they encounter when interacting with data and AI (RQ2); and the self-perceived opportunities of AI-supported reporting systems through the lens of discursive design (RQ3). Our findings reveal that local journalists do not fully leverage AI's potential to support data-related work. Despite local journalists' limited awareness of AI's capabilities, they are willing to use it to process data and discover stories. Finally, we provide recommendations for improving AI-supported reporting in the context of local news, grounded in the journalists' socio-technical perspective and their imagined AI future capabilities.

Figures (3)

• Figure 1: This figure presents our two operational prototypes: Figure \ref{['fig:prototypeA']}a, Automating-Through-Demonstration, appears in panels 1–3, whereas Figure \ref{['fig:prototypeB']}b, Automating-Through-Words, appears in panels 4–6. We did not include the city crime map scenario to preserve our anonymity during reviews, since it would have disclosed our location. In this scenario, the user wants to select all football teams listed on Wikipedia and identify their stadiums. The user can select all team names simultaneously by interacting with the table on the Wikipedia page. They can select all entries in the table in a single step. The system records these actions graphically and opens the first result/link. The user then selects the required text (2.). After demonstrating this once, the system can replay the actions continuously until it retrieves all the initially selected elements. The system generates an automation script in Typescript as shown in step (3.). The user can inspect and edit the script. The prototype shown in Figure \ref{['fig:prototypeB']} at step (4.) is our second, an automated browser that accepts user instructions through prompts. The crime map in step (6.) serves as a template that illustrates the information presented during interviews generated automatically after analyzing the data. In both prototypes, the user can examine the system logs and download the collected results as a JSON or an Excel File (5.).
• Figure : (a) Prototype 1. Automating-Through-Demonstrations
• Figure : (b) Prototype 2. Automating-Through-Words