Paper Type
ERF
Abstract
As Artificial Intelligence (AI) transitions from background automation to a front-stage "agentic partner," it reconfigures how users perceive and manage risk. We theorize the AI Guard-Down Effect: a sociotechnical failure mode where the delegation of interpretive labor to AI agents erodes human vigilance. We propose a model where AI-mediated interpretations act as authoritative "safety signals." These signals facilitate cognitive offloading, leading to Verification Displacement and a systemic reduction in threat appraisal. We present a multi-context experimental program designed to test these propositions across message triage, sensitive data sharing, and "Shadow AI" adoption. By identifying the mechanism through which "frictionless" AI design suppresses System 2 analytical processing, this research contributes to the Information Systems (IS) community’s understanding of trust calibration in adversarial environments. We conclude by proposing Strategic Friction as a design paradigm to restore human agency and mitigate the hidden costs of AI-human collaboration.
Paper Number
1288
Recommended Citation
Abouzahra, Mohamed, "The Mirage of Safety: Unpacking the "AI Guard-Down Effect" on Cybersecurity Vigilance and Privacy Behaviors" (2026). AMCIS 2026 Proceedings. 3.
https://aisel.aisnet.org/amcis2026/sig_sec/sig_sec/3
The Mirage of Safety: Unpacking the "AI Guard-Down Effect" on Cybersecurity Vigilance and Privacy Behaviors
As Artificial Intelligence (AI) transitions from background automation to a front-stage "agentic partner," it reconfigures how users perceive and manage risk. We theorize the AI Guard-Down Effect: a sociotechnical failure mode where the delegation of interpretive labor to AI agents erodes human vigilance. We propose a model where AI-mediated interpretations act as authoritative "safety signals." These signals facilitate cognitive offloading, leading to Verification Displacement and a systemic reduction in threat appraisal. We present a multi-context experimental program designed to test these propositions across message triage, sensitive data sharing, and "Shadow AI" adoption. By identifying the mechanism through which "frictionless" AI design suppresses System 2 analytical processing, this research contributes to the Information Systems (IS) community’s understanding of trust calibration in adversarial environments. We conclude by proposing Strategic Friction as a design paradigm to restore human agency and mitigate the hidden costs of AI-human collaboration.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
SIG SEC