The Third Moment of AI Ethics: Developing Relatable and Contextualized Tools

The "ethics gap" is real: While AI ethics literature and policy guidelines have proliferated, practitioners in industry still struggle to actually implement these principles in their work, suggesting that theoretical development has outpaced practical application.

Accessibility matters for adoption: The paper identifies that overly abstract language and poor accessibility are key barriers preventing AI ethics principles from being adopted by practitioners, indicating that how guidance is communicated is as important as what is communicated.

Practitioners need concrete, actionable tools: Rather than general principles, industry professionals require specific, step-by-step guidance integrated into their existing workflows (like software development processes) to effectively apply AI ethics in practice.

Participatory design with practitioners is essential: The researchers developed their tool by working directly with industry practitioners rather than imposing top-down solutions, suggesting that involving the people who must implement AI ethics is crucial for creating usable solutions.

Context-specific validation is important: The team validated their approach through a proof of concept study in the autonomous driving industry specifically, highlighting that AI ethics tools may need to be tailored to particular sectors and use cases rather than applied universally.