Defining diversity and inclusion in AI
We provide a normative definition for D&I in AI built upon the five pillars of the AI ecosystem:
Diversity and Inclusion in Artificial Intelligence refers to the ‘inclusion’ of humans with ‘diverse’ attributes and perspectives in the data, process, system, and governance of the AI ecosystem.
Diversity refers to the representation of the differences in attributes of humans in a group or society.
Attributes are known facets of diversity including the protected attributes in Article 26 of the International Covenant on Civil and Political Rights (ICCPR): race, color, sex, language, religion, political or other opinions, national or social origin, property, birth or other status; and (given the non-exhaustive nature of Article 26, attributes explicitly protected under Australian discrimination federal law, including but not limited to:) age, disability, criminal record, ethnic origin, gender identity, immigrant status, intersex status, neurodiversity, sexual orientation; and intersections of these attributes.
Inclusion is the process of proactively involving and representing the most relevant humans with diverse attributes who are impacted by, and have an impact on, the AI ecosystem context.
AI ecosystem refers to the collection of 5 pillars (humans, data, process, system, and governance), plus the environment (application or business domain) within which the AI system will be deployed and used.
A process diagram showing the application of Human, Data, Process, System and Governance elements to Diversity and Inclusion in Artificial Intelligence.
Artificial Intelligence Ecosystem process diagram
How to Navigate the Guidelines
The guidelines for Diversity and Inclusion in Artificial Intelligence are published as a chapter in the Book titled “Responsible AI: Best Practices for Creating Trustworthy AI Systems”.
We advocate that Diversity and Inclusion principles should be at the center of the AI ecosystem and must be embedded in all pillars for the specific context of AI use.
Below we present our guidelines by mapping them to the 5 pillars of diversity and inclusion in the artificial intelligence ecosystem as in our definition above. For each guideline, we provide suggestions for the potential job role of who should consider or be responsible for these guidelines. You can click on the job roles to find a short description and all the relevant guidelines for that specific job role.
