HQP: A Human-Annotated Dataset for Detecting Online Propaganda
Open accessAbstract
Online propaganda poses a severe threat to the integrity of societies.However, existing datasets for detecting online propaganda have a key limitation: they were annotated using weak labels that can be noisy and even incorrect.To address this limitation, our work makes the following contributions: (1) We present HQP: a novel dataset (N = 30,000) for detecting online propaganda with high-quality labels.To the best of our knowledge, HQP is the first large-scale dataset for detecting online propaganda that was created through human annotation.(2) We show empirically that state-ofthe-art language models fail in detecting online propaganda when trained with weak labels (AUC: 64.03).In contrast, state-of-the-art language models can accurately detect online propaganda when trained with our high-quality labels (AUC: 92.25), which is an improvement of ∼44%.(3) We show that prompt-based learning using a small sample of high-quality labels can still achieve a reasonable performance (AUC: 80.27) while significantly reducing the cost of labeling.(4) We extend HQP to HQP+ to test how well propaganda across different contexts can be detected.Crucially, our work highlights the importance of high-quality labels for sensitive NLP tasks such as propaganda detection.
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Still being checked.
Cited by
Sorted by citations
Feuerriegel, S., Hartmann, J., Janiesch, C., & Zschech, P. (2024) · Business & Information Systems Engineering · doi:10.1007/s12599-023-00834-7 · Cited by 1,409 · see citing works
“In addition, generative AI, specifically LLMs, can not only be used to spot harmful content on social media (e.g., Maarouf et al. 2023), but it can also create realistic disin-formation (e.g., fake news, propaganda) that is hard to detect by humans (Kreps et al. 2022; Jakesch et al. 2023).”
Self-citationUsing natural language processing to analyse text data in behavioural science
Feuerriegel, S., Maarouf, A., Bär, D., Geissler, D., Schweisthal, J., Pröllochs, N., Robertson, C. E., Rathje, S., Hartmann, J., Mohammad, S. M., et al. (2025) · Nature Reviews Psychology · doi:10.1038/s44159-024-00392-z · Cited by 140 · see citing works
Self-citationGenerative AIPreprint
Feuerriegel, S., Hartmann, J., Janiesch, C., & Zschech, P. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4443189 · Cited by 65 · see citing works
Self-citationChaudhari, D., & Pawar, A. (2023) · Big Data and Cognitive Computing · doi:10.3390/bdcc7040175 · Cited by 11 · see citing works
Together we can do it! A roadmap to effectively tackle propaganda-related tasks
Rodríguez-García, R., Centeno, R., & Rodrigo, Á. (2024) · Internet Research · doi:10.1108/intr-05-2024-0785 · Cited by 3 · see citing works
A Trope-Based Computational Approach to Participatory Propaganda Analysis
Niven, T., Li, C., & Lin, Y. (2025) · IEEE Transactions on Technology and Society · doi:10.1109/tts.2025.3592894
Butkienė, R., Šukys, A., Dambrauskas, E., Žitkus, V., Ablonskis, L., Vaičiukynas, E., Danėnas, P., & Butleris, R. (2026) · Applied Sciences · doi:10.3390/app16126251
- Authors
- Maarouf, A., Bär, D., Geissler, D., & Feuerriegel, S.
- Affiliations
- Munich Center for Machine Learning (DE)
- Venue
- Findings of the Association for Computational Linguistics ACL 2024
- Published
- 2024
- Pages
- 6064-6089
- OpenAlex
- W4402670741
- Cited by
- 7
- References
- 0
- Citations by year
- 2026: 1 · 2025: 2 · 2024: 1 · 2023: 3
- FWCI
- 5.24
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Findings of the Association for Computational Linguistics: ACL (Findings of ACL) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Deutsche Forschungsgemeinschaft
- Topics
- Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
- Physical Sciences › Computer Science › Artificial Intelligence › Hate Speech and Cyberbullying Detection
- Physical Sciences › Computer Science › Information Systems › Spam and Phishing Detection
- Keywords
- human-annotated dataset, weak labels, label noise, label quality, prompt learning
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.
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
