Responsible AI Guidance

The Responsible Al Checklist (RAIC) is intended to guide the development and evaluation of a complete Al solution and system against CHAI standards for trustworthy Al. This tool is intended first for self-reporting and self-review, as well as a tool for self-reporting for independent review.

The goal of the RAIC is to ensure that Al solutions and systems fulfill all five key, principle-based areas for trustworthy Al: 1. Usefulness, Usability, and Efficacy; 2. Fairness, Equity, and Bias Management; 3. Safety; 4. Transparency and Intelligibility; 5. Privacy and Security.

In alignment with these areas, the RAIC translates best practice considerations (detailed in the Responsible Al Guide) that meet core ethical and quality principles into detailed yes/no questions, or evaluation criteria, to determine whether best practice standards are met (see accompanying Responsible Al Guide).

The Responsible Al Checklist (RAIC) is intended to guide the development and evaluation of a complete Al solution and system against CHAI standards for trustworthy Al. This tool is intended first for self-reporting and self-review, as well as a tool for self-reporting for independent review.

The RAIC translates best practice considerations (detailed in the Responsible Al Guide) that meet core ethical and quality principles into detailed yes/no questions, or evaluation criteria, to determine whether best practice standards are met (see accompanying Responsible Al Guide).