As an artist, I would be trying to build/train a model off of my own work to help speed up the creative process and work with AI to create unique images that I couldn't think of.
Will you become more of yourself yesterday than yourself tomorrow?
No, because using AI for providing you with a starting point material trained on your own work (that you've already done before manually), and then building an artwork on top of that manually without any AI will let you spend less time on things you've already done (aka you-yesterday) and focus more on creating new things you haven't done yet (you-tomorrow). Which is the whole point.
Just because most visual art starts with basic shapes and builds everything on top of that doesn't make all that art "derivative" or "impeding your creative process".
An analogy to writing code, even though it isn't the best analogy - rewriting boilerplate starter code manually for a new project (instead of reusing the starter boilerplate code you wrote before for a previous project), isn't going to make your new code "more derivative" or make you "more of yourself yesterday". Doing so will let you simply spend less time and energy on the samey boilerplate and will let you focus on your actual novel code instead.
Artist style evolves over time, and it is influenced by outside sources, I agree with you. That evolution of style and influence will show up just fine in the manual parts of the work the artist is doing (past the ai-generated "starter boilerplate")
Aesthetics isn't just now grappling with the question of originality. This is an old conversation, people have been arguing that all creative work is technically derivative for (thousands of?) years.
To the parent’s question: on some kind of margin I think it would lead to more derivative work, but the net effect would still be novel creative output by a human. (I read the original idea as using the model as inspiration as much as a producer.)