Assume robots are the same as humans?

Suggests Nick.
We are unable to make that assumption (entanglement). The mere conversation assumes an observable difference. Instead assume that robots used up our quota of robot definition so that we could only have one definition of robot. Then that robot would be slightly different, the difference would be its ability to know things we cannot know, look where we cannot look, and use its slight bit of imprecision to define a new world. We and the robot sharing imprecision by entanglement.

Humans agree to a more precise near world, leaving the robots with some imprecision to define a far world, using Hanson language of near and far. In the language of polynomials, we define the near world coefficients and release the robots to find the polynomial coefficients that match a far world.