Implementing Ethics in AI: Initial Results of an Industrial Multiple Case Study

Ethics implementation in AI development remains largely informal: The study found that organizations lack formal methods and tools for implementing ethics in AI system design, despite growing recognition of the need for ethical considerations in AI development.

Healthcare sector shows limited structured ethical practices: Even in the high-stakes healthcare domain where AI decisions directly impact patient outcomes, companies are not systematically applying established ethical frameworks during development and deployment.

Gap between academic ethics literature and industrial practice: While conceptual AI ethics frameworks exist in academic literature, the research reveals a significant disconnect between these theoretical guidelines and what organizations actually do when building AI systems in real-world settings.

Stakeholder impact recognition is insufficient: The study indicates that while developers acknowledge AI systems affect multiple stakeholders, this awareness does not consistently translate into concrete ethical safeguards or stakeholder consideration during the design process.

Urgent need for practical implementation guidance: The findings suggest that organizations need more than just ethical principles—they require actionable methods, tools, and best practices that can be realistically integrated into existing software development workflows.