I've seen the Kaluza-Klein metric presented in two different ways., cf. Refs. 1 and 2.
In one, there is a constant as well as an additional scalar field introduced: $$\tilde{g}_{AB}=\begin{pmatrix} g_{\mu \nu}+k_1^2\phi^2 A_\mu A_\nu & k_1\phi^2 A_\mu \\ k_1\phi^2 A_\nu & \phi^2 \end{pmatrix}.$$
In the other, only a constant is introduced:
$$\tilde{g}_{AB}=\begin{pmatrix} g_{\mu \nu}+k_2A_\mu A_\nu & k_2A_\mu \\ k_2A_\nu & k_2 \end{pmatrix}.$$
Doesn't the second take care of any problems associated with an unobserved scalar field? Or is there some reason why the first is preferred?
References: