As discussions on inclusiveness often remain vague with respect to the term, it is impossible to assess whether the underlying “thing” is inclusive in the end. Let’s change that for inclusive AI.

While being intuitive, relevant and makes you look well informed, using the terms “inclusive” and “inclusiveness” is sometimes an escape from defining what level of equality or egality(?) shall be achieved.

I propose to attempt defining the term in a very early phase  - ideally by discussion with a large, diverse and open minded group. The ongoing discussions might over be time provide more insights and allow improvement of the definition.

Having said this, I is probably my obligation to present a first suggestion. Please add you suggestions, amendments and critique in the comments field. 

Inclusive means that a person’s access - to AI in this case - is not conditional on any explicit attribute that is outside the control of this person’s free will (or a combination of those) and any residual inequalities outside the person’s control are alleviated as far as possible.