Artificial Intelligence in Digital Agribusiness: A Content Analysis of Generative AI Skits and Agritech Narratives on TikTok and Instagram Reels

Abstract

Background: The rapid convergence of short-form video platforms and generative artificial intelligence has reshaped digital agribusiness communication. While short-form video formats offer unprecedented algorithmic reach, empirical scrutiny regarding the thematic framing, narrative structures, and technical accuracy of generative agritech content remains limited.

Objective: This study investigated how generative artificial intelligence skits and agritech narratives are framed, depicted, and received across TikTok and Instagram Reels, evaluating dominant thematic frames, narrative tones, technical feasibility, and audience engagement metrics.

Methodology: A multimodal quantitative content analysis was conducted on a sample of 400 publicly accessible videos (200 from TikTok and 200 from Instagram Reels) published between January 2024 and December 2025. Data were coded using a non-adapted coding sheet covering narrative tone, agritech framing, sub-sector, feasibility, and engagement metrics. Inter-coder reliability was established via Cohen kappa (ranging from 0.81 to 0.92). Statistical evaluation utilised chi-square tests of independence, Kruskal-Wallis H tests, and post-hoc Dunn tests with Bonferroni corrections.

Result: Advisory and farm management emerged as the dominant thematic frame (40.5%), followed by labour substitution (25.8%), concentrated largely in arable farming (41.0%) and livestock production (28.0%). Significant platform differences emerged: TikTok heavily favoured comedic skits and satire (67.0%), whereas Instagram Reels favoured educational and promotional formats (55.0%). Exaggerated or fictional AI capabilities accounted for 58.0% of the corpus, with realistic depictions representing only 42.0%. Comedic and satirical tones drove significantly higher audience engagement across views, likes, comments, and shares than instructional tones. Furthermore, exaggerated AI representations generated significantly higher virality than realistic workflows. Narrative format operated independently of the depicted agribusiness sub-sector.

Conclusion: Short-form algorithmic video platforms serve as influential cultural conduits for modernising agricultural perceptions; however, their algorithmic architectures systematically prioritise comedic satire and exaggerated technical capabilities over realistic agronomic education. This attention-driven dynamic risks fostering distorted perceptions of digital agribusiness feasibility among emerging agripreneurs, highlighting the necessity for grounded, institutional extension engagement on visual social platforms.