Beyond Hallucination: Semantic Laundering, Semantic Stewardship, and Prompting Aikido in Human-AI Collaboration
Abstract:
The primary conversation surrounding generative artificial intelligence centers on factual fabrication: hallucination. While hallucination matters, it represents a crude and easily detected failure mode. This paper identifies a second, quieter epistemic risk: semantic laundering. Large language models produce exceptionally fluent, coherent, and orderly prose that silently smooths away anomalies, unresolved tensions, weak signals, idiosyncratic language, and vital distinctions. In high-stakes domain inquiry, this polite normalization erodes the very irregular material on which professional judgment depends.
Drawing on a reflexive design case, this paper proposes semantic stewardship as a governing practice for preserving meaningful differentiation. It introduces prompting aikido: compact, context-sensitive micro-perturbations designed to nudge a model away from default attractors without over-specifying the output trajectory. The paper compares the human toolbelt (the Sideline Card) to an NFL call sheet, a human-operated instrument for rapid cognitive intervention rather than an automated script. While law serves as a high-friction exemplar, the framework applies universally across strategy, research, executive decision-making, and creative direction.
It also raises the core question: how do you prevent the machinery of compression from destroying the very distinctions expertise exists to notice?
#blogfirst
You can read and download the paper on Zenodo here: https://zenodo.org/records/22014530 or download the PDF directly from my website here. The paper is under a Creative Commons Attribution 4.0 International (CC BY 4.0) license to ensure the work remains open-access and free of proprietary barriers.
[Originally posted on DennisKennedy.Blog (https://www.denniskennedy.com/blog/)]
DennisKennedy.com is the home of the Kennedy Idea Propulsion Laboratory
Like this post? Buy me a coffee
DennisKennedy.Blog is part of the LexBlog network.