In the proof of the Radon-Nikodym theorem, the function $f = \frac{\mathrm{d} \nu}{\mathrm{d} \mu}$ is constructed as the supremum of measurable functions satisfying $\int_A f \mathrm{d} \mu \leq \int_A \mathrm{d}\nu$, and the existence of this supremum is by Zorn's lemma, an equivalent statement of the axiom of choice (AC), which is not admitted by many mathematicians.
So, does Radon-Nikodym's theorem still hold if we don't admit AC? And if it doesn't, then can we get a similar statement like Radon-Nikodym's theorem (probably weaker)?
You can actually prove versions of Radon-Nikodym without AC or LEM! Indeed, there are constructive (in the sense of Bishop) versions of the Radon-Nikodym theorem, though I haven't spent any time thinking about them, so I'm not sure exactly how much weaker they are, or which ~bonus assumptions~ you might need to make them work.
For instance, Bridges proves in The Constructive Radon-Nikodym Theorem that
According to the introduction of this paper, there is a different constructive version of Radon-Nikodym which can be found in Bishop's Foundations of Constructive Analysis (Ch 7. Thm 2), which applies only to measures on a locally compact space. I haven't looked at this reference, so it's possible that the restriction to locally compact spaces makes some other aspect of it more nice, which will work better for your purposes.
I hope this helps ^_^