Published July 12, 2026
| Version v1
Software
Open
Beyond Aggregate Sentiment: Machine Learning-Driven Discourse Indicators for AI News at Scale
Authors/Creators
-
1.
Jožef Stefan Institute
- 2. International Research Centre on AI under the auspices of UNESCO
- 3. SINTEF AS
Description
Code and derived results for Beyond Aggregate Sentiment: Machine Learning-Driven Discourse Indicators for AI News at Scale. A scalable DistilBERT pipeline extracts six theory-grounded discourse indicators: valence, loss-salience, narrative drift, exposure-adjusted sentiment, cross-source divergence, and novelty-phase framing, over 671,913 topic-linked AI-news headlines (386 topics), validated against GPT-4o and a three-annotator human benchmark, and mapped to EU AI-governance instruments.
Files
environment.txt
Additional details
Identifiers
Related works
- References
- Other: 10.5281/zenodo.15113095 (DOI)
Dates
- Submitted
-
2026-07-12
Software
- Repository URL
- https://github.com/E3-JSI/Beyond-Aggregate-Sentiment
- Programming language
- Python
References
- Alexiev, V., Bechev, B.& Graphwise/Ontotext. (2025). The InnoGraph Artificial Intelligence Taxonomy (Version v1.0) [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.15113095