P-E-A in the Digital Age: Towards Perception-Aware and Ethical Algorithm Design

In their early years, social media platforms were envisioned as spaces for diverse dialogue and discovery. In practice, however, users often encounter feeds dominated by extreme content and highly-engaged posts, reinforcing perceptions of consensus that diverge sharply from reality. While existing frameworks such as echo chambers, filter bubbles, and algorithmic bias partially explain these dynamics, they fail to quantify the perceptual distortion produced by engagement-driven amplification. To address this gap, this paper introduces Perceived Exaggerated Amplification (P-E-A) for the field of Human-Computer Interaction and social computing: a phenomenon in which platform algorithms systematically inflate the perceived prevalence and intensity of extreme or high-arousal content. P-E-A arises from three interacting mechanisms: algorithmic incentives that prioritize popular, emotionally-charged posts; human cognitive biases that disproportionately attend to salient or extreme content; and feedback loops in which repeated exposure reinforces the perception of widespread consensus. By operationalizing a P-E-A score—integrating engagement metrics, sentiment analysis, and network topology—platforms can detect perceptual distortion, evaluate algorithmic interventions, and guide ethical, human-centered system design. Empirical studies provide evidence for these dynamics. P-E-A offers a measurable, actionable lens for designers, engineers, and policymakers to reduce perceptual bias, improve digital literacy, and mitigate cognitive and social harm, enabling more accurate user perception and ethical platform governance in the digital age.