Time for AI (Ethics) Maturity Model Is Now

Principles alone are insufficient: While governments, public bodies, and universities have published numerous AI ethics principles in recent years, the paper finds that these high-level principles do not easily translate into actionable guidance that practitioners can actually implement in their work.

AI development needs a structured improvement roadmap: The paper advocates for adopting "maturity models"—a proven approach from software engineering that provides companies with clear stages and key capabilities for progressively improving their performance, rather than expecting organizations to implement ethics perfectly from the start.

AI ethics requires a software engineering perspective: The authors argue that AI systems should be treated as software development challenges, not just ethical philosophy problems, suggesting that established software engineering practices and frameworks are essential for managing AI responsibly.

The scope question matters: The paper raises an important debate about whether the focus should be narrowly on "AI ethics" or more broadly on overall "AI system quality," suggesting that a comprehensive maturity model might need to address multiple dimensions of responsible AI development beyond ethics alone.

Companies need practical guidance, not just declarations: The observation that companies are now publishing their own ethical guidelines indicates a gap between what high-level principles promise and what organizations actually need—structured, step-by-step pathways to improve their AI development practices.