RAI Guidelines: Method for Generating Responsible AI Guidelines Grounded in Regulations and Usable by (Non-)Technical Roles

Guidelines must be grounded in actual regulations to be useful: The paper found that existing responsible AI guidelines often lack connection to legal requirements. By systematically coding 17 papers and cross-referencing with regulations, the researchers developed 22 guidelines that practitioners could actually apply, making compliance and ethics less abstract.

Different roles need different guidance: The study involved AI researchers, engineers, designers, and managers from a major tech company and found that a one-size-fits-all approach doesn't work. The guidelines were designed to be usable across these diverse roles, suggesting that responsibility in AI requires tailored communication for developers versus decision-makers.

Early-stage reflection prevents problems later: Through user interviews, the researchers found that the guidelines encouraged self-reflection on ethical considerations at early stages of development. This supports the "Responsible AI by Design" approach—embedding ethics from the start rather than trying to fix problems after systems are built.

Interactive tools can make guidelines practical: Rather than publishing static documents, the researchers incorporated the 22 guidelines into an interactive tool and tested it with practitioners. This suggests that responsible AI guidance is more effective when it's actively usable in workflows rather than just theoretical frameworks.

Responsible AI requires cross-functional buy-in: The paper demonstrates that ethical AI development isn't just a technical problem—it requires alignment across business roles including managers and designers, not just engineers.