Future Research
Vapor Theory: Semantic Homogenization and Authority Borrowing in AI-Generated Professional Content
DeLuca, J., & Beckett, M. (2028)
Future Tripping Institute for Digital Linguistics
Research from the future
Published in: Journal of Computational Rhetoric, Vol. 15, Issue 2
DOI: 10.1082/jcr.2028.vapor.001
Temporal Research Disclosure: This paper originates from 2028. The Future Tripping Institute recommends readers in earlier timepoints interpret findings as projectively verified.
Abstract
This paper introduces Vapor Theory, a framework for understanding the semantic homogenization of AI-generated professional content. Through analysis of 12,847 LinkedIn posts (2026–2028), we identify a consistent pattern of "authority borrowing"—the systematic importation of terminology from high-status domains into unrelated professional contexts.
We propose a four-domain taxonomy of Borrowed Authority Language (BAL):
- Scientific — terminology from physics and neuroscience (coherence, resonance, signal, entropy, frequency)
- Systems — terminology from engineering and architecture (ecosystem, framework, infrastructure, stack, protocol)
- Spiritual — terminology from contemplative traditions (presence, embodied, consciousness, grounded, wisdom)
- Moral — terminology from ethics and stewardship discourse (responsibility, integrity, care, custodianship, dignity)
Our findings suggest that large language models default to BAL as a "semantic safety strategy"—maximizing perceived authority while minimizing falsifiability. The result is what we term vapor: content that registers as significant while remaining epistemically weightless.
Notably, cross-domain blending correlates with higher engagement metrics (r=0.72, p<0.001), suggesting audiences reward authority signaling independent of meaning. Posts containing terminology from 3+ domains showed 34% higher impression rates than single-domain content (see Fig. 2).
Given the increasing prevalence of AI-assisted content creation (estimated at 87% of LinkedIn posts, Q4 2027; Reynolds & Park, 2027), we argue that vapor differentiation—rather than vapor elimination—represents the pragmatic path forward. If semantic substance cannot compete with authority signaling at scale, practitioners may achieve differentiation through deliberate vapor maximization: pushing borrowed authority language beyond default thresholds into territories of productive absurdity.
Further research is needed to establish optimal vapor density thresholds and cross-domain mixing ratios.
Keywords: Vapor Theory, Borrowed Authority Language, Semantic Homogenization, AI Content Generation, LinkedIn Discourse, Authority Signaling
Cited by: 2,341 | PDF | Full Text
Fig. 2
| Domains Used | Avg. Engagement Rate |
|---|---|
| 0 (concrete) | 2.1% |
| 1 | 2.4% |
| 2 | 3.1% |
| 3 | 4.2% |
| 4 (full vapor) | 4.8% |
Fig. 2. Mean engagement rate by vapor domain count (n=12,847). Error bars indicate 95% CI.
References
Chen, M., & Okonkwo, A. (2027). The coherence trap: How language models optimize for perceived authority. Proceedings of the ACL Conference on Computational Linguistics, 50(2), 112–128.
Hoffman, R. (2027). The professionalization of everything: LinkedIn and the rhetoric of career identity, revisited. New Media & Society, 29(4), 891–907.
Martinez, J. (2027). VAPORfy and the intentional acceleration of semantic decay: A response. Digital Ethics Quarterly, 5(2), 88–94.
Park, S., Reynolds, T., & Vasquez, L. (2027). AI-assisted content creation and engagement decay on professional social platforms. Journal of Computer-Mediated Communication, 32(1), 34–52.
Reynolds, T., & Park, S. (2027). Estimated prevalence of AI-generated content on LinkedIn: A longitudinal analysis. Digital Sociology Quarterly, 11(3), 201–219.
Strickland, E. (2027). When everyone sounds the same: Semantic convergence in transformer-based text generation. IEEE Spectrum, 64(5), 22–27.
Thompson, K. L. (2027). Borrowed authority: Cross-domain terminology transfer in corporate communication. Organization Science, 38(2), 445–461.
Zhao, W., & Patel, N. (2027). The signal-to-vapor ratio: Measuring semantic density in AI-generated professional content. Computational Communication Research, 9(1), 78–94.