AI-Ethics by Design. Evaluating Public Perception on the Importance of Ethical Design Principles of AI

The public views ethical AI principles as roughly equally important, rather than prioritizing one over others. In a survey of 1,099 German respondents evaluating AI use in tax fraud detection, participants weighted explainability, fairness, security, accountability, accuracy, privacy, and machine autonomy comparably, suggesting society expects comprehensive ethical design rather than shortcuts.

One-size-fits-all ethical AI design is unrealistic because the population divides into distinct preference clusters with substantially different priorities and importance levels for ethical principles. This means developers cannot satisfy everyone's ethical concerns equally, requiring deliberate choices about which groups' values to prioritize.

Demographic and ideological differences shape what people consider ethically important in AI systems. The study found that sociodemographic characteristics and people's existing opinions about AI significantly influenced which ethical principles they valued most, indicating that ethical AI design cannot ignore diverse public perspectives.

Trade-offs between ethical principles are unavoidable in practice. The researchers note that simultaneously meeting all ethical principles is "not only costly, but sometimes even impossible," meaning developers must make explicit choices about which ethical values to prioritize—decisions that should ideally reflect public input rather than being made unilaterally.

Public input on ethical AI design is urgently needed but currently lacking. Despite AI ethics being framed as "human-centric" and beneficial for society, this appears to be one of the first studies examining what the public actually thinks matters ethically in AI systems, highlighting a gap between who AI affects and who decides its ethical standards.