Navigating the Grey Zone: Estonian Communication Practitioners’ Perspectives on Ethical AI Use and Responsibility in Crisis Communication

Sammanfattning

Artificial intelligence (AI) is increasingly employed in crisis communication to assist with tasks like monitoring, analysis, and message creation. However, its expanding use raises ethical and responsibility concerns, especially when AI-supported communication results in problematic outcomes. This thesis aims to provide new knowledge and enhance the understanding of how Estonian communication professionals perceive the ethical use of AI and responsibility in crisis communication. Using semi-structured, vignette-based interviews with 16 practitioners from 13 organisations and drawing on moral responsibility and machine ethics, the study explores how practitioners define AI’s role, where they set boundaries between human and AI tasks, and how they consider responsibility when AI-related errors occur. The analysis indicates that practitioners view AI as a support tool but recognise that it can subtly influence their interpretation of crises. Factors such as crisis severity, the ability to verify outputs, and data sensitivity primarily guide the division of tasks between humans and AI. Regarding moral responsibility, practitioners do not see it as a single issue: crisis conditions can diminish judgements of individual competence, organisational hierarchy can influence who is held answerable for explaining failures regardless of who used the AI, and accountability for outcomes remains unclear. The thesis highlights that practitioners using AI operate in a highly uncertain environment, where few widely agreed-upon norms or established practices guide their actions. It therefore underscores the need for organisational AI guidelines and structured evaluation procedures to better support practitioners, who currently face AI-related challenges largely on their own.