Epistemic Possibility and Textual Source Voice in AI Technology Journalism: A Corpus-Assisted Case Study of May, Might, Could, and Likely, 2022–2026
Journal of Northern Europe Academy for Studies and Research · 2026
DOI: 10.53348/nea3se33 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Northern Europe Academy for Studies and Research, 2026. Verfasst von Asjad Balla.
Abstract (Originalsprache)
AI journalism often couples statements about current technical capability with predictions whose realization is uncertain. This exploratory, corpus-assisted case study investigates may, might, could and likely in 20 criterion-based TechCrunch AI articles, chosen as two eligible articles from two fixed 14-day windows each year, 2022–2026. On October 4, 2026 a script was executed to audit the captured canonical pages, raw html, cleaned editorial text and verification logs. Case-insensitive, whole-word searches in headlines and editorial bodies yielded 40 target tokens. Was was the most frequent form (n = 16, 40.0%) and occurred in nine articles, but only 7 of the 16 occurrences (43.8%) were epistemic; the other nine were capacity or affordance. Epistemicity was found in 10 of the 12 may tokens (83.3%) and all the might and likely tokens. Of these, 29 (72.5 %) were epistemic. One watermarking article in 2023 accounted for 8 of the 40 tokens, so removing it dropped the proportion of all target tokens in 2023 from 22.5% to 3.1%. Two blinded machine-assisted coding passes agreed on 35 of 40 semantic codes (87.5%; κ = .679). This is an algorithmic audit diagnostic, not human inter-rater reliability. The findings are descriptive, entirely auditable, and limited to this small, purposively bounded corpus Keywords: epistemic modality, uncertainty, artificial intelligence, technology journalism, corpus-assisted discourse studies,modal verbs, source attribution
Zitieren
APA
Balla, A. (2026). Epistemic Possibility and Textual Source Voice in AI Technology Journalism: A Corpus-Assisted Case Study of May, Might, Could, and Likely, 2022–2026. Journal of Northern Europe Academy for Studies and Research. https://doi.org/10.53348/nea3se33
Vancouver
Balla A. Epistemic Possibility and Textual Source Voice in AI Technology Journalism: A Corpus-Assisted Case Study of May, Might, Could, and Likely, 2022–2026. Journal of Northern Europe Academy for Studies and Research. 2026. doi:10.53348/nea3se33
BibTeX
@article{Balla2026,
author = {Balla, Asjad},
title = {Epistemic Possibility and Textual Source Voice in AI Technology Journalism: A Corpus-Assisted Case Study of May, Might, Could, and Likely, 2022–2026},
journal = {Journal of Northern Europe Academy for Studies and Research},
year = {2026},
doi = {10.53348/nea3se33}
}
EUVIMED