Benedikt Graf , Rebecca Müller , Thomas Ellwart , Conny Herbert Antoni
{"title":"More or less? Effects of different levels of automation of a software agent on information overload, workload, and stress perceptions","authors":"Benedikt Graf , Rebecca Müller , Thomas Ellwart , Conny Herbert Antoni","doi":"10.1016/j.chbr.2025.100727","DOIUrl":null,"url":null,"abstract":"<div><div>While high automation of software agents aims to reduce workplace demands such as workload and information overload, software agents are often perceived as a source of technostress rather than a resource. Current theoretical models emphasize the positive impacts of automation but neglect the influence on technostress and human stress perception, leaving the mechanisms behind stress reduction by software agents unexamined. In this pre-registered study, we investigated how automation of software agents reduces stress and technostrain and if reduced workload and information overload mediate these effects. A randomized within-experiment was conducted in which 108 humans worked with a software agent with four different automation levels in randomized order and rated overload, stress, and technostrain after each automation level. ANOVA and mediated multilevel structural equation model results showed that information overload (η<sup>2</sup> = .55, <em>p</em> < 001), workload (η<sup>2</sup> = .71, <em>p</em> < 001), stress (η<sup>2</sup> = .63, <em>p</em> < 001), and technostrain (η<sup>2</sup> = .48, <em>p</em> < 001) decreased with higher automation of software agents. Mediating processes indicated that stress as well as technostrain decreased via the reduction of workload (β = −.56, <em>p</em> < .001), but not due to the reduction of information overload (β = .06, <em>p</em> = .331). This study expands the knowledge on how automation and autonomy of software agents can be perceived as a resource instead of a techno-stressor and integrates classical models of stress, overload, and human autonomy research. Moreover, this study shows that organizations can use automation of software agents to reduce technostrain.</div></div>","PeriodicalId":72681,"journal":{"name":"Computers in human behavior reports","volume":"19 ","pages":"Article 100727"},"PeriodicalIF":5.8000,"publicationDate":"2025-06-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Computers in human behavior reports","FirstCategoryId":"1085","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2451958825001423","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"PSYCHOLOGY, EXPERIMENTAL","Score":null,"Total":0}
引用次数: 0
Abstract
While high automation of software agents aims to reduce workplace demands such as workload and information overload, software agents are often perceived as a source of technostress rather than a resource. Current theoretical models emphasize the positive impacts of automation but neglect the influence on technostress and human stress perception, leaving the mechanisms behind stress reduction by software agents unexamined. In this pre-registered study, we investigated how automation of software agents reduces stress and technostrain and if reduced workload and information overload mediate these effects. A randomized within-experiment was conducted in which 108 humans worked with a software agent with four different automation levels in randomized order and rated overload, stress, and technostrain after each automation level. ANOVA and mediated multilevel structural equation model results showed that information overload (η2 = .55, p < 001), workload (η2 = .71, p < 001), stress (η2 = .63, p < 001), and technostrain (η2 = .48, p < 001) decreased with higher automation of software agents. Mediating processes indicated that stress as well as technostrain decreased via the reduction of workload (β = −.56, p < .001), but not due to the reduction of information overload (β = .06, p = .331). This study expands the knowledge on how automation and autonomy of software agents can be perceived as a resource instead of a techno-stressor and integrates classical models of stress, overload, and human autonomy research. Moreover, this study shows that organizations can use automation of software agents to reduce technostrain.
虽然软件代理的高度自动化旨在减少工作场所的需求,如工作量和信息过载,但软件代理通常被视为技术压力的来源,而不是资源。目前的理论模型强调自动化的积极影响,但忽视了对技术压力和人类压力感知的影响,使得软件代理减少压力背后的机制未经研究。在这项预先注册的研究中,我们调查了软件代理的自动化如何减少压力和技术紧张,以及减少的工作量和信息过载是否介导了这些影响。在一项随机的内部实验中,108个人按照随机顺序与一个具有四种不同自动化水平的软件代理一起工作,并在每个自动化水平后评估过载、压力和技术应变。方差分析和中介多层次结构方程模型结果显示,信息过载(η2 = 0.55, p <;001),工作量(η2 = 0.71, p <;001),应力(η2 = 0.63, p <;001),技术应变(η2 = 0.48, p <;001)随着软件代理自动化程度的提高而降低。中介过程表明,压力和技术应变通过工作量的减少而减少(β =−)。56、p <;.001),但不是由于信息过载的减少(β = .06, p = .331)。本研究扩展了软件代理的自动化和自主性如何被视为一种资源而不是技术压力源的知识,并整合了压力、过载和人类自主性研究的经典模型。此外,本研究表明,组织可以使用软件代理的自动化来减少技术压力。