AI in the newsroom - A data quality assessment framework for employing machine learning in journalistic workflows

Aug 3, 2026

AI-driven journalism covers methods and tools for gathering, verifying, producing, and distributing news information, aiming to extend human capabilities and enable augmented journalism. While accountability requires embedding journalistic values, data quality receives less attention, even though accuracy and efficiency depend on high-quality data. Defining data quality is multidimensional and domain-dependent, so assessment needs an interdisciplinary approach treating journalists as end-users and balancing machine-learning and ethical challenges. This work proposes a conceptual framework to support collection and pre-processing, strengthening data literacy and highlighting limitations and biases.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages9

Reading Time23 min

AccessFree to download

Introduction

Theoretical backdrops

5th International Conference on Advanced Research Methods and Analytics (CARMA2023) Universidad de Sevilla, Sevilla, 2023 DOI: [http://dx.doi.org/10.4995/CARMA2023.2023.16440](http://dx.doi.org/10.4995/CARMA2023.2023.16440) AI in the newsroom: A data quality assessment framework for employing machine learning in journalistic workflows Laurence Dierickx1, Carl-Gustav Lindén1, Andreas L Opdahl1, Sohail Ahmed Khan1, Diana Carolina Guerrero Rojas1 1Department of Information Science and Media Studies, University of Bergen, Norway Abstract AI-driven journalism refers to various methods and tools for gathering, verifying, producing, and distributing news information. Their potential is to extend human capabilities and create new forms of augmented journalism. Although scholars agreed on the necessity to embedjournalistic values in these systems to make AI-driven systems accountable, less attention was paid to data quality, while the results' accuracy and efficiency depend on high-quality data. However, defining data quality remains complex as it is a multidimensional and highly domain-dependent concept. Assessing data quality in AI-driven journalism requires a broader and interdisciplinary approach, considering journalists as end-users. It means meeting the challenges of data quality in machine learning and the ethical challenges of using machine learning in journalism. These considerations ground a conceptual data quality assessment framework that aims to support the collection and pre-processing stages in machine learning. It aims to strengthen data literacy in journalism by emphasizing limitations and possible biases related to data and making abridge between journalism studies and scientific disciplines that should be viewed through the lenses of their complementarity. Keywords: data quality assessment, journalism, ethics, machine learning, artificial intelligence This research was funded by EU CEF grant number 2394203 . This work is licensed under a Creative Commons License CC BY-NC-SA 4.0 Editorial Universitat Politcnica de Valncia 217 1. Introduction AI-driven journalism refers to various methods and tools for news gathering, verification, production, and distribution (Thurman et al., 2019) . They aim to support professional practices to help speed up time-consuming tasks, publish automated content, identify trends, or provide insights into large numeric or textual datasets. Hence, their potential is to extend human capabilities and augment journalism practices (Lindén, 2018) . Although AI-driven systems are often considered opaque and not bias-free (Guidotti et al., 2019), they depend on high-quality data to avoid inaccurate analytics and unreliable decisions (Gupta et al., 2021) . Explaining how data is collected, organised, cleaned, annotated, and processed participatesin establishing a relationship of trust between the journalist as the end-user and the tool. It implies understanding the challenges of data quality that appear upstream and downstream of a machine learning process (Gudivada et al., 2017) . The “garbage in, garbage out” principle also applies in journalism, whereas quality information requires quality data to ensure the accuracy and reliability of the news (e.g., Anderson, 2018; Diakopoulos, 2020; Dierickx, 2017; Dörr & Hollbuchner, 2017; Lowrey et al., 2019). However, less attention was paid to this critical aspect. The conceptual framework presented in this paper intends to fill this gap, considering that assessing data quality is context and use dependent (Tayi & Ballou, 1998; Boydens & Van Hooland, 2011) . 2. Theoretical backdrops Data quality encompasses several complementary dimensions referring to a set of attributes in which dimensions – such as accuracy, completeness, and consistency – were refined overtime. However, research agreed that data quality refers to data that adapts to the uses of data consumers, especially in terms of accuracy, relevance, and understandabil

FAQs

Why is data quality crucial for AI-driven journalism workflows?

AI-driven tools rely on high-quality data to avoid inaccurate analytics and unreliable decisions. Data quality also affects trust between journalists and the end-user systems.

What makes assessing data quality in journalism difficult?

Data quality is multidimensional and domain-dependent, so it must adapt to the intended use and consumer needs. The concept also becomes more complex with big data and diverse structured and unstructured sources.

How does the paper propose to address data quality during machine learning?

It presents a conceptual data quality assessment framework designed to support collection and pre-processing stages. The framework emphasizes limitations and possible biases and connects journalism studies with scientific perspectives through complementarity.

Aug 3, 2026

Upload by Olivia Brown