1) This part of the question holds even without requiring $AB=BA$ or $A^2=B^2$. Let $T$ be a selfadjoint operator that commutes with $A-B$. Then, for any vector $v$ we have $Pv=(A-B)w$ for some $w$, so
$$
TPv=T(A-B)w=(A-B)Tw.
$$
This implies that $TP=PTP$. As we were assuming $T=T^*$, we have $PT=(TP)^*=(PTP)^*=PT^*P=PTP=TP$. So $T$ commutes with $P$. If you now consider a general $T$, it is written as a linear combination of selfadjoints (i.e. real and imaginary parts), and the result follows.
2) This one is not true. Let
$$
A=B=\begin{bmatrix}1&0\\0&0\end{bmatrix},\ x=\begin{bmatrix}0\\1\end{bmatrix}.
$$
Then $AB=BA$, $A^2=B^2$. We also have $A-B=0$, and so
$P=0$. Then
$$
Ax=0,\ Px=0\ne x.
$$
3) Since $AB=BA$, the operators $A$ and $B$ can be simultaneously diagonalized, i.e. there exist spectral measure $E$ and Borel functions $f,g$, which are zero off the support of $E$, such that
$$
A=\int\,f\,dE,\ \ B=\int\,g\,dE.
$$
From $A^2=B^2$ we get that $f^2=g^2$. Then there exists measurable $h$, taking values in $\{-1,1\}$, such that $g=hf$. So
$
A-B=\int\,(1-h)f,
$
and $P=E\{h=1\}$. So
$$
2P-I=2\int_{h=1}\,1\,dE-\int\,1\,dE=\int_{h=1}],1\,dE-\int_{h=-1}\,1\,dE=\int h\,dE.
$$
So, noting that $h=h^{-1}$,
$$
(2P-I)B\int hg\,dE=\int\,f\,dE=A.
$$