Fig 1.
Semantic landscapes of cognitive labor: The shift from process inquiry to product generation.
(a) AI prompts (Upper) are dominated by imperative verbs (answer, write, give) and output specifications (text, sentence), reflecting an “Instrumental Director” stance. (b) Human questions (Lower) are grounded in procedural nouns (Method, Problem, Way) and evidence markers, reflecting a “Struggling Petitioner” stance grounded in friction.
Fig 2.
Syntactic markers of role displacement.
(a) AI interactions (Left) prioritize ontological definition (“You are...”) and direct command (“Give me...”), establishing immediate epistemic authority. (b) Human interactions (Right) rely heavily on autobiographical narratives (“I have...”, “I am trying...”) to contextualize the request and validate user effort.
Fig 3.
The structural dissolution of defensive social norms.
(a) Distribution of Hedges: Human interactions (Blue/Right) maintain a normative spread, whereas AI interactions (Orange/Left) show absolute convergence at zero. (b) Distribution of Politeness Markers: The contrast between the “mandatory baseline” in human groups and the “zero-inflation” in AI groups confirms the removal of the “Politeness Tax” in algorithmic environments.
Fig 4.
Variance in self-focus as an indicator of social deregulation.
The compact Interquartile Range (IQR) of the HHI group (Blue) indicates tight social norms constraining self-presentation. In contrast, the elongated IQR of the HAI group (Orange) reveals a polarization between “silent commanders” (zero self-focus) and “confessional users” (high self-focus), characteristic of an environment lacking social supervision.
Fig 5.
The internal consistency of impression management strategies.
(a) The HAI group (Left) exhibits small but detectable positive correlations among social markers, reflecting a slight behavioral consistency in the absence of external constraints. (b) The HHI group (Right) displays near-zero correlations (“white noise”), indicating that impression management is a “norm-driven” strategic module deployed only when socially required.
Fig 6.
The distribution of cognitive vulnerability disclosure.
The Human group (Blue) displays a “performative humility” tail (x > 0), reflecting the social necessity of admitting ignorance to legitimize help-seeking. The AI group (Orange) exhibits a leptokurtic peak at zero, confirming an “Instrumental Directness” where learners bypass the admission of cognitive gaps.
Fig 7.
Affective dynamics and emotional labor.
The HHI distribution (Blue) shows a high-positivity tail (x > 0.75), quantifying the “facework” required to maintain community relations. The HAI distribution (Orange) is anchored at neutral (0.0), illustrating the affective sterility of frictionless inquiry.