Delegating Decisions, Avoiding Responsibility: The Ethical Risks of Organizational AI

The study examines the ethical implications of delegating organizational decisions to artificial intelligence. It investigates how AI reliance, AI autonomy, and governance clarity relate to responsibility shifting, moral disengagement, and unethical decision intentions. The research combines an experimental design with a three-wave survey to examine responses to organizational AI under controlled conditions and associations across successive measurement occasions. Study 1: Experimental Research The supplied Study 1 dataset contains 1,712 records and 40 variables. It includes three binary experimental factors: AI reliance, AI autonomy, and governance clarity. Together, these factors represent eight possible experimental conditions. The dataset includes manipulation-check items, measures of responsibility shifting, moral disengagement, unethical decision intentions, and a behavioral-choice variable. Background information includes age, gender, managerial role, industry, and years of AI experience. Participant codes identify individual records. The research documentation must specify the recruitment procedure, eligibility requirements, allocation method, experimental materials, screening criteria, and final analytical sample. The meaning and coding of the behavioral-choice variable must also be documented. Study 2: Three-Wave Survey The supplied Study 2 dataset contains 675 records and 63 variables. It includes participant codes, cohort information, dates for three measurement occasions, and a matched-completion indicator. These fields support the identification and linkage of responses across survey waves. The dataset contains measures of AI reliance, AI autonomy, AI governance clarity, responsibility shifting, moral disengagement, and unethical decision intentions. Additional variables cover job autonomy, time pressure, ethical climate, marker items, and other questionnaire items. Background information includes gender, age group, organizational position, industry, AI-use frequency, workplace AI experience, weekly working hours, organizational tenure, managerial responsibility, and decision-task category. The research documentation must identify which measures were administered at each wave, the intervals between waves, initial recruitment numbers, attrition, matching procedures, exclusions, and the final analytical sample. Data Protection and Participant Welfare Consent materials must explain the experimental or survey activities, follow-up contact arrangements, voluntary participation, foreseeable discomfort, and confidentiality protections. Where responses are linked across waves, participant codes and any identifying contact information must be stored separately with restricted access. Individual responses concerning workplace conduct and ethical decision-making must remain confidential and must not be disclosed to employers or supervisors. Data retention, access, sharing, and any international transfers must follow the approved protocol and applicable institutional requirements. Approval Coverage and Research Records The ethical approval must identify the study components, participating institutions, recruitment locations, participant groups, and data collection procedures it covers. Coverage of multiple countries or institutions must be supported through the original approval or a documented amendment. Ethical approval, project affiliation, and preregistration are separate records. Any preregistration claim must identify a verifiable timestamped record and accurately report its timing and subsequent deviations. DATA PROTECTED AND COPYRIGHT BY SRI (UIBE).