Introduction: While existing research has largely emphasized technical and legal aspects of Artificial Intelligence (AI) applied to policymaking processes, the human dimension—attitudes, perceptions, and acceptance of relevant stakeholders —remains fragmented and underexplored.
Purpose: To synthesize empirical evidence on how data-driven AI tools are perceived, adopted, and evaluated by relevant stakeholders, such as citizens and policymakers, within public policy-making processes. Results will advance the understanding of psychological and social factors in AI data-driven policy making processes.
Methods: A PRISMA-compliant systematic literature on experimental, quasi-experimental, and qualitative studies retrieved from Scopus and Web of Science before 1st July 2025 (pre-registered (protocol: osf.io/zwgr7). From 1232 records screened, 58 studies met the inclusion criteria.
Results: Barriers such as limited digital skills, organizational constraints, inadequate algorithmic understanding, and concerns over privacy, security, and trust were identified. Facilitators included stakeholder involvement in design and decision-making, transparent communication of AI capabilities, and governance frameworks stressing fairness, accountability, and ethics. Psychological factors were central: trust emerged as the main predictor of acceptance, mediating the effects of transparency, perceived usefulness, and privacy, while fairness perceptions, resistance to change, and job insecurity also shaped attitudes.
Conclusions: These findings provide the first comprehensive synthesis of attitudes toward data-driven policy tools. Beyond the policy perspective, they offer applied psychology concrete entry points for intervention—such as designing training and change management programs, developing trust-building and communication strategies, and addressing job insecurity and resistance to change—to foster responsible and psychologically sustainable AI integration in the public sector.