Definition
Textual Inflation
The rapid increase in cheap, fluent, AI-generated text that reduces the marginal signaling value of individual text outputs while increasing filtering, verification, attention and trust costs.
Evidence for Attention Scarcity under Generative Abundance
Large language models have changed the economics of textual production by making fluent, plausible prose inexpensive to generate at scale. This paper develops the concept of textual inflation: the rapid increase in cheap, fluent, AI-generated text that reduces the marginal signaling value of individual text outputs while increasing filtering, verification, attention and trust costs.
The paper synthesizes prior work on information overload, cognitive load, attention scarcity, the attention economy, calm technology, glanceable interfaces, AI slop, AI fatigue, generative search, model collapse and authenticity concerns.
When generated text increases faster than available human attention and verification capacity, filtering burden and authenticity uncertainty rise, making curation, provenance, trust signals and compressed semantic forms more valuable.
The central question is whether large-scale LLM-generated textual abundance creates measurable pressure toward semantic compression: lower-bandwidth meaning carriers such as summaries, icons, reactions, badges, provenance marks, confidence displays, dashboards, haptics, ambient cues and other glanceable signals that preserve enough relevance, state, trust or intent while reducing attention cost.
The paper does not argue that text disappears, nor that color or chromatic systems are necessary solutions. Its claim is narrower: LLM-era textual abundance increases the value of trust, provenance, curation and compressed semantic forms.
The rapid increase in cheap, fluent, AI-generated text that reduces the marginal signaling value of individual text outputs while increasing filtering, verification, attention and trust costs.
The degradation of distinctiveness, trust, nuance or interpretive stability under conditions of excessive symbolic output.
The leftover cognitive burden created by verification, filtering, rewriting, tool switching and uncertainty management.
The use of lower-bandwidth meaning carriers that preserve relevance, state, trust or intent while reducing attention cost.
The tendency for users, platforms and interfaces to prefer lower-friction meaning carriers when text becomes too abundant.
Semantic compression supports text by routing attention toward the right depth of explanation, rather than replacing language itself.
| Theory or discourse | What it explains | What TSX-6 adds |
|---|---|---|
| Information overload | Excessive information overwhelms limited processing capacity. | Defines textual inflation as a generative-text-specific overload condition. |
| Attention economy | Attention becomes scarce and economically valuable under abundance. | Connects cheap generated text to declining marginal signaling value and rising curation value. |
| Cognitive load theory | Working memory is limited and excessive information creates cognitive burden. | Links verification, rewriting and uncertainty management to cognitive residue. |
| Calm technology | Interfaces can inform through peripheral, low-friction cues. | Treats calm and glanceable forms as semantic compression under LLM-era pressure. |
| AI slop discourse | Synthetic saturation creates banality, sameness, clutter and trust decline. | Places AI slop within a structural model of textual inflation and compression pressure. |
| Model collapse | Recursive synthetic training can degrade distributions under some conditions. | Provides an adjacent system-level analogue for semantic entropy and human-signal scarcity. |
Repeated prompting, supervising, evaluating and integrating generated output can create cognitive strain.
Fast AI-generated outputs may look polished while shifting evaluation and repair burdens to recipients.
Mass-generated synthetic content increases the effort required to distinguish signal from generic output.
Origin, authorship and perceived human contribution affect value judgments, increasing demand for provenance.
AI summaries can reduce source engagement while increasing the importance of citation and source visibility.
Synthetic feedback loops create a system-level analogue for semantic entropy and human-signal scarcity.
The notation is not presented as a proven mathematical law, but as a map for operationalizing future research.
For human-computer interaction, TSX-6 suggests that text-heavy interaction may not remain optimal for every AI-mediated task. As LLMs produce more text, the interface problem shifts from generating language to managing attention.
Better support may mean producing less text, offering clearer status, showing uncertainty, preserving provenance and compressing state. Agentic systems may need dashboards, state indicators, action logs, reversible summaries and trust cues rather than long prose after every action.
Text is not disappearing. It remains one of the most powerful media for reasoning, explanation and institutional memory. The TSX-6 claim is more specific: when generated text becomes abundant, the marginal signaling value of fluent text declines while the value of attention, trust, curation, provenance and compression rises.
LLMs make text easier to produce, but they do not make human attention easier to expand.