CoChat

A Taxonomy of LLM-Supported Tasks for Critical AI Literacy in Journalism

Open access
Dierickx, Laurence, Bjerknes, Fredrik, Opdahl, Andreas L., Lindén, Carl-Gustav · 2026 · Postdigital Science and Education

The DOI matches its Crossref record; 64 of the 64 references it cites were found in a registry or an index; none of the 58 sources its claims cite is checked for support yet.

Abstract

This paper develops a task-based taxonomy for integrating Large Language Models (LLMs) into journalism education, drawing on task definitions from machine learning and natural language processing alongside journalism studies. Organised around four stages of the news workflow, newsgathering, sensemaking, editing, and publication/distribution, it translates journalistic activities into computationally informed tasks linked to baseline prompts and ethical risks. The taxonomy is designed to help students use LLMs critically while maintaining editorial standards. It addresses key risks, including bias, hallucinations, and overreliance on automation. More broadly, the approach promotes AI literacy in journalism by combining structured practice with ethical reflection and emphasising the importance of human oversight in generative AI–assisted workflows. While grounded in journalism, the taxonomy’s underlying logic is transferable to other knowledge-intensive professions where accuracy, source evaluation, and accountability are at stake, offering a conceptual basis for developing similar taxonomies in other postdigital professional training contexts.

Overview

CoChat Summary

TL;DR

The paper proposes a task-based taxonomy for integrating LLMs into journalism education across four news-workflow stages—newsgathering, sensemaking, editing, and publication/distribution—linking each task to baseline prompts and ethical risks. It argues that structured task definition and prompting can foster critical AI literacy and human oversight in generative AI-assisted journalism.

Key findings

  • Develops a task-based taxonomy organised around four stages of the news workflow: newsgathering, sensemaking, editing, and publication/distribution.
  • Maps each stage to computational tasks (e.g., classification/clustering, detection/discovery, extraction, linking, summarisation; comparison, evaluation, exploration, prediction, prioritisation; augmentation, correction, generation, segmentation, simplification; adaptation, conversion/translation, generation, personalisation, reformulation) and pairs them with baseline prompts and risk/mitigation strategies.
  • Frames prompting as a situated journalistic practice and task definition as professional judgement, addressing risks including bias, hallucinations, and overreliance on automation while emphasising human oversight.
  • States three contributions: a workflow-based taxonomy, an ethical framework linking LLM use to journalistic values such as accuracy and accountability, and a conceptualisation of prompting as situated journalistic practice; it also claims transferability to other knowledge-intensive professions.

Methods

  • Integrative approach combining synthesis of journalism studies on news work and epistemic practices, task definitions from ML/NLP literature, and professional experience teaching AI-related methods in journalism education and newsroom environments.
  • Candidate task categories derived from ML/NLP and journalism studies literature and iteratively refined against newsroom practice using abductive logic; categories were merged, split, or relabelled.
  • No datasets were generated or analysed; the taxonomy is a practice-oriented pedagogical tool with intentionally simple, model-agnostic prompts.

Limitations

  • The paper states the taxonomy is not exhaustive and requires validation in classroom settings; it should be regarded as a pedagogical resource rather than a prescriptive model.
  • Evident from the text: no empirical classroom intervention or validation data are reported, so claims about improving students’ critical evaluation remain to be tested.
  • Evident from the text: prompts and risk categories are designed to be transferable across tools and versions, but their effectiveness across contexts is not empirically assessed.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1007/s42438-026-00690-0

Cites

64 of 64 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Verification

Claims the paper makes

0 Valid · 1 Not valid · 42 Not checked · 39 Own results and interpretations

10 more claims

Create a free account to see all 82 claims and to verify the 39 references still open.

See all claims — free account
  • Probably not a substantial claim. 60
  • It may state two separate claims as one. 18

Create a free account to see these claims.

Citations

Cites

Show all 64 references

Research works printed but not linked: 7

  1. Behzadi, F. (2015). Natural language processing and machine learning: A review. International Journal of Computer Science and Information Security, 13(9), 101–106.
  2. Biswas, M., & Bland, D. (2024). AI in journalism and communication teaching: Use, challenges and opportunities. Teaching Journalism and Mass Communication, 14(1), 11-20.
  3. Cai, M., & Nishal, S. (2023). Motivations, goals, and pathways for AI literacy for journalism. In CHI’23 Workshop on AI Literacy: Finding Common Threads between Education, Design, Policy, and Explainability.
  4. Ettema, J. S., & Glasser, T. L. (1998). Custodians of Conscience: Investigative Journalism and Public Virtue. New York: Columbia University Press.
  5. Joseph, S. R., Hlomani, H., Letsholo, K., Kaniwa, F., & Sedimo, K. (2016). Natural language processing: A review. International Journal of Research in Engineering and Applied Sciences, 6(3), 207-210.
  6. Li, C., & Diakopoulos, N. (2024). Probing GPT-4 for knowledge of journalistic tasks. In Proceedings of Computation + Journalism Symposium (C+J Symposium). https://cplusj2024.github.io/papers/CJ_2024_paper_17.pdf. Accessed 2 August 2026.
  7. Tuchman, G. (1978). Making News: A Study in the Construction of Reality. New York: Free Press.

Cited by

See all 2

Metadata

Authors
Dierickx, L., Bjerknes, F., Opdahl, A. L., & Lindén, C.-G.
Affiliations
University of Bergen (NO)
Corresponding author
Dierickx, Laurence
Venue
Postdigital Science and Education
Published
2026-08-18
ISSN
  • 2524-485X (ISSN-L)
  • 2524-4868
OpenAlex
W7203723387
Cited by
2
References
62
Citations by year
2026: 2
FWCI
8.04
Type
journal article
Language
en
Open access
Licence
Locations
Postdigital Science and Education (publishedVersion, cc-by-nc-nd)
Funders
University of Bergen
Topics
  • Social Sciences › Communication › Journalism and Media Studies
  • Social Sciences › Education › Artificial Intelligence in Education
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
Keywords
large language models, AI literacy, journalism education, news gathering, sensemaking, editorial editing, AI hallucination, algorithmic bias, over-reliance on automation, human oversight, generative AI, ethical risks of AI, source evaluation, accountability

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-20, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-20 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: A Taxonomy of LLM-Supported Tasks for Critical AI Literacy in Journalism. https://app.cochat.ai/papers/10.1007/s42438-026-00690-0
The paper
Dierickx, L., Bjerknes, F., Opdahl, A. L., & Lindén, C.-G. (2026). A Taxonomy of LLM-Supported Tasks for Critical AI Literacy in Journalism. Postdigital Science and Education. https://doi.org/10.1007/s42438-026-00690-0

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.