The inverted spectra problem isn't a problem for the example I gave.

The hypothetical humans in the problem do not exist.

Real, flesh and blood humans exist.

Even if two real humans have isomorphic relationships with color, and try to paint the same thing, they'll make choices in how they use color. When creating art, not making choices is not an option. Their subjective experiences mean they won't make the same choices.

They'll choose colors in different orders. They'll mix paints differently. There will be minute motor differences. They'll perceive something, translate it to their own inner world, and then transport it out again via fine motor skills & paint.

In the final product, they -won't- have produced the same work of art, because their subjective humanity ensured that their processes would not be isomorphic. At the same time, they will have communicated details of their inner subjective experience, in an objective fashion--using a known medium. Even if you train the artists, or the elephant artists, this process is going to happen.

We've defined objective reality via consensus--and the sheer body of evidence surrounding the average experience of what redness is, is well-established. It can -feel- different from person to person & this difference can be readily communicated.

Nagel's Bat is a hypothesis designed to be untestable & immune to evidence.