A systematic literature review published in Artificial Intelligence Review examined the operational impact of artificial intelligence systems deployed in high-stakes environments where human lives are at risk. Researchers Nataly R. Panczyk and Majdi I. Radaideh evaluated the safety risks, protective mitigations, and governance structures surrounding automated and assisted decision-making across safety-critical domains.
The investigation addressed an operational gap where rapid advancements in artificial intelligence technologies outpace the regulatory frameworks designed to protect the public. While prior reviews examined adjacent computational questions, the authors identified an absence of literature syntheses focusing specifically on safety-critical industries and the structural relationships connecting technology developers, operational end-users, and regulatory authorities.
To capture relevant research, the investigators conducted a broad database search across Web of Science and Scopus on December 15, 2025. The initial literature retrieval yielded an evidentiary pool of 1,135 candidate papers evaluating automated and algorithmic decision-making tools.
The research team applied targeted screening criteria to evaluate each candidate system’s operational sensitivity, decision-making framework, and associated risk profile. This screening narrowed the initial collection to a final corpus of 72 reports, each subjected to formal quality assessment and risk of bias evaluations to verify methodological rigor.
From the analyzed reports, the authors identified critical system characteristics and industry deployment practices organized into three primary areas: the safety risks introduced by utilizing artificial intelligence in critical choices, practical mechanisms to mitigate those hazards, and procedures for assigning accountability and responsibility for computer-generated decisions.
The review extracted both categorical and binary variables from the accepted studies to establish quantitative patterns regarding the specific types of artificial intelligence implemented, the technical expertise of operational end-users, and the decision-making time windows allotted to human operators, complemented by per-industry qualitative assessments of operational trends.
Published on November 9, 2026, as an open-access work in Artificial Intelligence Review, the study was released in an early citable form with a permanent digital object identifier to facilitate rapid access to peer-reviewed findings ahead of the final Version of Record.
Subject of Research: Artificial intelligence in safety-critical decision-making
Article Title: When computers call the shots: a systematic review exploring the impact of AI in high-stakes decision-making
Article References: Panczyk, N. R., & Radaideh, M. I. (2026). When computers call the shots: a systematic review exploring the impact of AI in high-stakes decision-making. Artificial Intelligence Review. https://doi.org/10.1007/s10462-026-11734-z
Image Credits: AI Generated
DOI: 10.1007/s10462-026-11734-z
Keywords: artificial intelligence, systematic literature review, high-stakes decision-making, safety-critical systems, risk mitigation, algorithmic accountability, technology regulation, decision support, computers, call, shots, systematic
News Source: Blake Davidson. (November 9, 2026). Systematic Review Examines AI Risks and Accountability in High-Stakes Decision-Making. Scienmag.



